| WIKI | FAQ | Tech FAQ | |
 http://csua.com/feed/ | |
| ||||||
| 5/16 |
| 2002/1/7-8 [Science/Space, Computer/HW/CPU] UID:23479 Activity:low |
1/7 In the spirit of the post below regarding 0.1 (IEEE floating point
representation), how do I find out what the rounding rule is for
a particular processor, and what DBL_EPSILON is? Would this pose a
big problem for scientific applications (e.g. Mars Rover, satellite,
atomic energy simulations, etc)?
\_ DBL_EPSILON is defined in /usr/include/float.h. I don't know if
the value is only for C running on that processor or if it's good
for any language running on that processor.
\_ Go talk to Kahan. Be prepared to be insulted for your stupidity.
But hey, it's a small price to pay.
\_ Apparently, Kahan can be neutralized if you talk to him in
the Right Way. For instance, a friend of mine would go ask
him a small question about linear algebra, then he just sits
back for the next hour and listen about linear algebra fairy
tales with dragons of perturbation, maidens of symmetric
forms, and knights of floating points. He highly recommends
it to anybody. Unfortunately I think this Right Way involves
being humble and perceptive, hence unsuitable for most of this
\_ Precisely. And I highly recommend the experience if
you have time to kill (having tried it
first-hand). -alexf
\- prepare to be quizzed. also Kahan's
insults/derision is pretty funny. it's
always "sincere" rather than "mean-
spirited" if you know what i mean.
--partha "meromorphic" banerjee
audience. Too bad. |
| 2002/1/6 [Computer/HW/CPU, Computer/HW] UID:23478 Activity:nil |
0.1 cannot be accurately represented in the binary format, and if
so, how come adding 0.1 successively on my computer works?
$PREC=64;
$radix=1;
$target=0.1;
print("0.");
for ($i=0; $i<$PREC; $i++) {
$radix/=2;
if ($radix<=$target) {
$target-=$radix;
print("1");
} else {
print("0");
}
}
print("\n");
\_ have you taken 61c? IEEE-754 floating point doesn't use all 64 bits
for the significand, you know. And no, 0.1 (decimal) can't be
represented exactly in binary. Try it on paper.
\_ the point of the poster is that since it cannot be represented
exactly in binary, why it works on modern processors.
\_ Well we know that certain Intel cpus just ignore all that
precision stuff anyway. :-) |
| 2001/12/27-28 [Computer/HW/CPU] UID:23381 Activity:nil |
12/26 Confused about Celerons, Durons, Athlons, K6-2, K7s?
This may help:
http://www.mkdata.dk/click/hardware/module3e01.htm |
| 5/16 |
| 2001/12/20 [Computer/HW/CPU] UID:23326 Activity:high |
12/19 Looks like Itanic is sinking in the harbor:
http://www.eweek.com/article/0,3658,s%253D1884%2526a%253D19288,00.asp
With problems like this and Hammer just around the corner
Intel's as good as dead!
\_ Chuckle. You fail to account for intertia.
\_ Let's see, P4, RamBus, Home Electronics (keyboards, mices,
homerf) and Commercial Networking (enterprise switches)
were all disasters. Itanic is late, slow and buggy and
runs Windows very badly. They are dying, and Hammer will
hammer the nails into Intel's coffin.
\_ You count P4 as a failure? It's a cash cow, man! If you
think AMD is going to kill Intel, then put your money down
and sell Intel short. Or buy AMD. I own some AMD but I'd
never sell Intel short. You're crazy. Intel isn't going
away anyime soon just because AMD has a new chip coming.
\_ P4 is a failure, it isn't selling nearly as well as
expected and people still want to buy P3s because the
performance on P4 is really really bad and you have
to use that stupid Rambus pos. Intel has made all bad
moves in recent years, this coupled with Hammer are
going to put Intel under. - AMD stock holder
\_ Uh whatever. Don't bet the farm on AMD fellow AMD
shareholder. Intel sure as hell isn't going under. |
| 2001/12/11-12 [Computer/HW/CPU] UID:23207 Activity:high |
12/07 Here's some benchmarks for same video/different CPUs:
http://www.tech-report.com/reviews/2001q4/vidcards/index.x?pg=11
\_ in another word, the quality is so fuckin good nowadays that the
nekked eyes can't really tell the difference between a $100 and a
$300 card/CPU unless you use benchmarks. Cool.
\_ Cool. I just bought a $100 card.
\_ for raw polygon, probably not much. But then there is
anti-alias, motion-compensation, hypertexture and other
which would bog down a GeForce3. The real question is
where are the games which use these features?
\_ Nothing out right now that I know of. |
| 2001/12/10-11 [Computer/HW/CPU] UID:23201 Activity:high |
12/9 I'm a gamer. Should I upgrade my CPU/motherboard to AMD or Pentium
(same clock rate)? I already have a kick ass video card. ok thx
\_ Price/perfomance ratio-wise, Athlon. AMD 1.4Ghz = Intel 2.0Ghz
in terms of performance. http://www.anandtech.com for more info.
\_ http://tomshardware.com did a cpu/gpu speed comparison a few months ago
which showed that an AMD Athlon 800 was just as good as the 1400
when you had a top end video card. No more FPS from the higher
speed cpu since the video card was the bottle neck. I'd check this
out before spending anything. I have a much older system so if I
was to buy today I'd get an Athlon 1600 or so but if I already had
anything 800 or higher I wouldn't bother.
\_ The problem is, most places don't carry anything that "slow"
anymore. Yeah, they're still out there, but with 1.4Ghz now
under $150, why bother going for anything less?
\_ All I'm getting at is that if the poster is already running
800+ then they may not see *any* performance increase. I'm
running p3-450 so I would but they *may* not. Without
knowing what they're upgrading from it's hard to say if
an upgrade is a waste of money or not. But, yes, sure, for
a new system from a much older system like mine, yeah I'd
go 1600+ just because the price difference is so minimal
anyway.
\_ It's sad that you had to basically repeat your original
post slightly reworded because the dumbass above didn't
understand what you were saying in the first place.
He must have been a stubborn [Arab|Jew|Democrat|Republ]. |
| 2001/11/4-5 [Computer/HW/CPU] UID:22932 Activity:high |
11/4 Does anyone know where I might be able to get a Athlon
mobo that can be setup via the serial port? I'd like to
get a Athlon but I want one that I can install and setup
without having to get a VGA monitor and PS2 keyboard.
\_ It'll cost a bit more (probably bringin the price up to around
that of Intel serial-control mobo's), but just add:
http://www.realweasel.com --dbushong
\_ This thing rocks. Thanks.
\_ I ordered three of their PCI cards a couple months ago. They're
still not ready yet (by which I mean that there have been delays
in bringing them to market, not just with my order). I hear the
ISA version is really available.
\_ Their FAQ and product pages say they're shipping PCI. You
might want to give them a call and remind them you exist. |
| 2001/10/10 [Computer/Domains, Computer/HW/CPU, Computer/HW/Display] UID:22680 Activity:nil |
10/10 Is is my network or are others having problems with
http://soda.berkeley.edu. I can't ping, can't access email the last
2 days. http://csua.berkeley.edu works.
\_ it's you. what does http://soda.berkeley.edu have to do with not being
able to access your email? soda = soda.csua = csua
\_ well i thought so too. But soda.b.e and csua.b.e map
to different ip's.
\_ It is 4:30am and the IST servers still report the old DNS info
because they either have not been updated or the changes take a while
to propagate. Also, if your ISP's DNS server cached old data, you might
see odd things happen even after the changes propagate through IST DNS
servers. Other places might already see the correct DNS info if they
happen to use the OCF DNS server because it has already been updated to
account for CSUA name server IP change. If it was up to me I'd do this
change in a two step process, leaving the name server running with
the old IP address for a few days so that at least the correct
information for other CSUA computers would have time to propagate.
\_ I have been having problems, too, but it is not because of DNS.
Even with the correct IP address from the DNS server, I cannot
log in and my traceroutes die at a http://cs.berkeley.edu router. |
| 2001/10/10 [Computer/HW/CPU, Computer/SW/OS/Windows] UID:22675 Activity:moderate |
10/9 what's the big deal about athlonXP? why are they calling it XP?
\_ marketing bs with M$
\_ they deny this of course.
i think it should have been called
"Super Athlon XP Professional Turbo Champion Edition"
\_ And what about "Athlon 4". Did I miss Athlon 2 and Athlon 3? ;p
\_ weren't those the CPUs that Netscape 5 was optimized for?
\_ only the Solaris 6 version.
\_ AMD is not competing on MHz. But they still need something catchy to
show that they are the latest (and thus implicitly the best). Since
Windows XP is marketed by MS to be the latest and best, the "Athlon
XP" trademark invokes the images of the latest and newest (and thus
the best). Pretty smart actually, since MS will pay for all the
marketing, and AMD doesn't have to follow the MHz war as tightly. |
| 2001/9/24 [Computer/HW/CPU, Computer/SW] UID:22610 Activity:nil |
9/23 how do you log a bitchx session into a file?
\_ you can run bitchx from within a screen session.
\_ In epic /set log on |
| 2001/9/17-18 [Computer/HW/CPU] UID:22492 Activity:nil |
9/17 http://dailynews.yahoo.com/h/zd/20010917/tc/rambus_intel_sign_new_chip_deal_1.html "The companies renegotiate their patent cross-licensing deal and announce a new 5-year agreement." Didn't Intel announce that they were going to break all ties with Rambus a while ago? \_ something on http://tomshardware.com about intel wanting some other memory controller stuff rambus has. I didn't read the full article. |
| 2001/9/10 [Computer/HW/CPU, Computer/SW/OS/Linux] UID:22362 Activity:moderate |
9/9 What the hell is wrong with you people? LinuxWorld
just passed and there was NOT ONE MENTION of it on
the motd! Don't you know that i COUNT on the motd to
keep me informed of geek stuff like this?
\_ Every second exhibitor was a hardware company. The biggest booths
were those of IBM (biggest and most visible), Compaq, HP and Sun
along with Intel and AMD. Dell and SGI were not there. Plus, there
were RedHat, Mandrake, SuSE, VaLinux/OSDN, Ximian, Debian, FSF and
other usual suspects. Unlike before, there wasn't much free gear
like CD's and t-shirts though you could pickup lots of magazines
and get massage for free. There wasn't much new this year, really.
Probably slashdot and such have better coverage of this event..
\_ Who cares about Linux (and LinuxWorld)?
\_ lisux |
| 2001/8/23-24 [Computer/HW/CPU] UID:22230 Activity:high |
8/23 INTC is going to SUCK. -anonymous coward
\_ INTC being?
\_ The stock symbol of Intel Corp.?
\_ I dunno, 2 GHz @ ~$550 per processor in bulk sounds pretty non-suck
to me. |
| 2001/8/23-24 [Computer/HW/CPU] UID:22220 Activity:nil |
8/23 so why is that apple doing all the PowerPC evangelism? Is it just
because they are the only company using PowerPC for personal computers?
shouldn't the company/ies that make the PowerPC be doing some of their
own evangelism too? Something to keep up with "Intel Inside"?
\_ Apple is the only consumer PC companies with PowerPC computer.
Definitely, they are up against the Pentium <insert number here>
more than the embedded and workstation market. You don't see
IBM marketing as heavily in the RISC market because people
who pay >$20k for a computer knows their stuff. On the other
hand, the new Nintendo Gamecube will have an IBM logo to
advertise for IBM/PowerPC. (Don't remember seeing a MIPS logo
on the Playstation) |
| 2001/8/20 [Computer/HW/CPU] UID:22185 Activity:insanely high |
8/20 Anybody know what is the focus of UC Merced? I've been going through
their web site and they talk a lot about engineering. But hell, so
does DeVry. How are they going to attract the famous professors and
get the reputation as a good school? I don't think that many people
are moving to Merced/Fresno. Thanks.
\_ the focus is the destruction of california's largest vernal pool.
to build that site.
\_ consider how many people move to Riverside and Davis.
\_ Davis is tons better than Riverside
\_ Top tier: B, LA, SD. Mid: SB, I, D. Bottom: SC, R
\_ SB vs D vs R is not a given. Why do you place R below the
other two?
\_ League of it's own: UCB.
\_ not if we graduate people who can't distinguish
its/it's.
\_ How do the low-tier UCs compare to the Cal States?
\_ Big money to get the prof's, and many years of time to gradually
build the reputation. Hmm, if they can also run a biz school to
attract the gals, I think more guys will be willing to go there
for engineering.
\_ [readded post]. Harvey Mudd got a good reputation in <50 yrs.
\_ just as a side note: Harvey Mudd's CS is ass.
\_ Their focus is going to be agricultural stuff, as expected
by its location. It was a lousy idea to create another UC.
\_ how do you suggest UC deal with expanded enrollment, then?
-tom
\_ don't enroll so many f'n people!
\_ Exactly. nuff said. no more of this nonsense
"let's admit the top 4%" plan.
\_ Competing with UCD? |
| 2001/8/13-14 [Computer/HW/CPU] UID:22096 Activity:high |
8/13 First Apple fought the MHz Myth, now its AMD's turn:
<DEAD>netscape.zdnet.com/zdnn/stories/news/0,4586,5095403,00.html<DEAD>
\_ Unfortunately, most people in this country have never taken a
class like CS 152 and just can't grasp this concept. As a
result most people believe that a chip with higher clock speeds
is automatically faster than one with a lower frequency. What's
worse is that Steve Jobs actually said something right for once
making the "MHz myth" less credible.
\_ People who care about cars know the difference between HP
and torque. People who care about computers will understand
the important criteria for CPUs. Don't sweat it.
\_ not to be jaded, but, whoop-pe-de-doo-dah
\_ You dont need to take cs152. You just need to understand that
AMD Athlons or IBM/Motorola PPC's get more done each cycle than
an Intel Pentium-whatever. It can use less cycles to get the
same amount of work done, and by using less cycles, use less
an Intel Pentium-whatever. It can use fewer cycles to get the
same amount of work done, and by using fewer cycles, use less
power -- saving the environment, increasing the lifetime of the
chip, and saving the skin of your lap from being burned off by
a wasteful overheated CPU.
\_ That's 'fewer' cycles, you dumbass. Geez, no wonder they
make engineers take E190.
\_ The name is Dumass.
\_ fine so I wrote this w/o the morning caffeine. does more
with less work. more efficient. good enough peon?
\_ My Celeron 400MHz at home is almost twice as slow as my P-II
350MHz at my previous job. Why can't AMD run an ad to say that MHz
is just like RPM on a car engine, that it's not the RPM that counts
but how fast a car can go at the given RPM that counts? One engine
rev on a 6-cyliner can move the car more inches than one engine rev
on a 4-cyl, so a 6-cyl car can be going faster than a 4-cyl car at
a particular moment even though the RPM on the 4-cyl is higher at
that time. One difference between CPUs and engines is that CPUs
runs at a constant clock rate, so the ad may need to compare the car
speeds with the engines at their red lines, assuming that they can
find two cars that are capable to reach their red lines at their
top gears and the red line on the 6-cyl is lower than that on the
4-cyl. --- yuen
\_ Have you used both CPUs on -identically- configured systems?
\_ No. --- yuen
\_ People have been suggesting this very idea to sjobs@apple.com
for several years. I think the main problem with it is that
the average pc buyer has no idea what RPMs are. The only possible
good that will come of this idea (if implemented) is that people
will want cars with more RPMs and Honda stock holders will benefit.
\_ Oh, I forgot that most people drive automatic these days.
--- yuen
\_ "Making life more difficult is always better!"(tm)
\_ That's why they invented computers, man.
\_ My Celeron 300 (oc'd to 450) is faster than my PII 400Mhz... |
| 2001/8/6-7 [Computer/HW/CPU] UID:22019 Activity:nil |
8/6 Someone asked about intel price cuts. Aug 26
http://news.cnet.com/news/0-1003-200-6792635.html?tag=mn_hd |
| 2001/7/24-25 [Computer/HW/CPU, Computer/SW/Unix/WindowManager] UID:21933 Activity:insanely high |
7/24 Is there a version of, or an equivalent to, the CDE (common
desktop environment) for linux?
\_ there's the KDE...
\_ Can it be made to look similar to CDE? All the
screen shots I've seen on the web look more like Windows.
screen shots I've seen on the web look more like
Windows.
\_ http://kde.themes.org, there used to be a cde theme.
\_ Look for TriTeal's XMotif/CDE package. Redhat used to
sell it as a add on to RedHat Linux up until v5.2 or
so. You can probably still get it running on newer
versions by using the 2.0 kernel/libc5 compat rpms.
---ranga
\_ TriTeal went out of business. http://www.xig.com is the
only company who sells CDE for Linux now. -alan-
\_ Why in god's name would you want a shitty interface
like cde? There are much better interfaces available
for *nix: twm, fvwm, afterstep, saw{mill,fish}, etc.
\_ Trade up from Linux to Solaris - it's still free and
includes CDE.
\_ In addition to CDE, you get *WORKING* NFSv3, PThreads,
SMP, MultiThreaded TCP/IP, RealTime Scheduling
Extensions and a stable kernel interface layer at
no additional charge! Who could pass up a deal like
that?
\_ I work with x86 daily. all i gotta say is:
"nfs server mamba not responding."
"nfs server pasteur not responding."
There's a reason why solx86 and staroffice are free. - paolo
There's a reason why solx86 and staroffice are
free. - paolo
\_ paolo, that is because Instructional is constantly
underfunded and short on resources appropriate to
the demands placed on them by coursework. I believe
this was in the undergraduate presentation made to
the department at the faculty retreat recently. --Jon
\_ All I gotta say is: "Incompetant NFS/network
\_ Wow, you think the architecture makes NFS work poorly?
Impressive.
admins"
\_ Solaris/sparc is just as free as the solaris
for x86.
\_ Poster never said he was on x86. He could be
running sparc linsux for all we know.
\_ Its your own damn fault for using cheap x86
hardware for NFS. Get a real machine.
\_ Wow, you think the architecture makes NFS work
poorly? Impressive.
\_ no, but the components that go into a x86
system are usually sub-par creating all
sort of problems. (Cheap raid cards, cheap
nics, cheap drives, 33 MHz bus, etc).
In anycase people who depend on NFS should
be using a dedicated solution like netapp
instead of running on a cheap pc they slapped
together from cheap bits at frys.
\_ as opposed to Sun's low-end server, the
extremely high-quality Ultra 10? Please.
-tom
\_ The U10 is not a low end server. It is
a desktop machine. If you buy a U10 and
think that its a server that's your own
damn fault for being cheap. A "low-end"
server would be the discontinued U30 or
U60...
\_ gee, you'll have to mention that to Sun,
which sells the Ultra 10 as a server on
their web site. If you're going to pay
$10K for a box, you'll get a much better
PC box than Sun box--if you buy shitty
PC hardware, "that's your own damn fault."
-tom
\_ Due to lack of funds, our Solaris/x86 NFS
server is a cheap PC thrown together from
parts. We haven't had any problems with it
so far.
\_ Who is talking about cheap x86 hardware?
You can buy a decent x86 server with 66MHz PCI
slots if you need, hot swapable SCA drives, and
other components. It will be faster than and just
as reliable as an equivalent Sun (or whatever)
machine but cost you two times less or more.
Netapps are nice but the cost is an issue with
them. For a price of a netapp filer you can buy
a disk array from Sun with at least the same
storage capacity and a server to go with it.
\_ So does your x86 "server" support gigE
fiber and etherchannel? How about ATM?
\_ Yes. You haven't looked at current Dell, IBM
etc offerings, have you?
Let's not forget about a real journalling
fs and seamless ha fall over. What about
NEBS compliance? Lights out management?
People who depend on NFS require these
things. People like you who play with
toys in your dorm room don't and think
that cheap x86 stuff from fry's is just
as good.
With regards to Netapp vs. Sun. Netapp
is faster and more reliable than a
multipurpose server. The WAFL fs is
probably the most advanced journaling
fs ever written and these boxes even
commit metadata updates to nvram so
that a powerfailure won't corrupt the
fs. If I depended on NFS I'd buy netapp.
\_ export list limit of 800.
can we say CIFS/Unix perm conflicts?
mountd dying periodically
I can go on. Netapp's are amazing
and I like things like wafl, snapshots,
the dual platform support, but they are
far from perfect. be realistic. --Jon |
| 2001/7/21 [Computer/HW/CPU] UID:21896 Activity:nil |
7/20 I just put together this new system. Tyan motherboard 266 fsb,
1 ghz athlon thunderbird, 768 mb ram, etc with 300 watt atx
power. case one starts athlon fan but no power to the board nor
system. case two, which is known to work, powers the fan
(meaning in both situations that when i plug the power cable in
the athlon fan briefly turns on) but still no power to the
system. is my motherboard fried? if so why does the fan turn
on? why can't i get any response from any components besides
the fan? am i forgeting to turn something on in the
motherboard?
\_ Could be the memory. I've seen this on dual proc machines with
the wrong type of memory.
\_ thanks for restoring the legit post.
\_ maybe you need more power? aren't Athlons power-hungry? shouldn't
you have a 400+ W power supply?
\_ does your fan plug into the motherboard or directly?
\_ I had strange problems with a P/S on an Athlon box, with
symptoms similar to yours. An original P/S apparently died
in transit; a working P/S from a P3 system wouldn't bring the
system up -- the motherbord didn't like it. Eventually managed
to get an Athlon-approved P/S for testing, the system finally
woke up. Drove to CompUSA and got a $100 P/S with the highest
power rating (400+W) -- working almost non-stop for 6+ months.
But it was a very frustrating experience. |
| 2001/7/16-17 [Computer/HW/CPU] UID:21813 Activity:high |
7/16 I tried the mips3-sgi-irix6.2 and the mips4-sgi-irix6.2 versions of
SETI@home on the same IRIX64 machine, and the latter only gives about
5% higher speed over the former. Anyone knows why a 64-bit version of
such data-intensive program as SETI@home doesn't give much higher
performance over its 32-bit counterpart?
BTW, my 250MHz R10000 IRIX seems to be able to complete a work unit in
about the same time as my 1.2GHz P4 NT. Wow!
\_ Why would it? A 64-bit architecture would only save you on a
few instructions as you attempt to load the register file with
64-bit values. That's about it. So even if SETI was using
64-bit it wouldn't run much faster. SIMD extensions like AltiVec
and MMX can derive some sub-word parallelism but MIPS isn't a
SIMD extension. 64-bit versions of IRIX also have a newer ABI
(n64) which could theoretically be more efficient (with register
allocation rules and such) but I'm sure it'snot significant
\_ But wasn't the selling point of 32-bit apps/CPUs over 16-bit
apps/CPUs in the industry years ago that "your data-intensive
applications will run much faster it can process thirty-two bits
of data at a time instead of sixteen bits"? The same doesn't
apply to 32bit --> 64bit? I understand that there are certain
instructions that doesn't benefit from a bigger data size (like
testing a boolean), but I was still expecting a difference bigger
than 5%.
\_ Probably because 16-bit programs attempted to mimic a 32-bit
computation by breaking it up into two parts and merging
the results together somehow. Plus, in the x86 world you
were limited to 2^16 bytes which meant that if you had
programs larger than 64K it would require that you use the
segment:offset addressing which isn't very efficient itself. |
| 2001/7/14 [Computer/HW/CPU, Computer/HW/Drives] UID:21797 Activity:nil |
7/13 Slashdot's starting to get pretty lame, is there anything else
like it?
\_ There are many things like it. This is, of course, the problem.
\_ I like http://arstechnica.com. Far fewer articles, but more interesting.
And the people who run the site are generally more intelligent and
make more informed comments, unlike CmdrTaco.
\_ More hardware oriented: http://www.anandtech.com
\_ http://www.theregister.co.uk
\_ The Register can be as witty as The Onion but it's all true.
\_ http://www.newsforge.com --Galen |
| 2001/7/12 [Computer/HW/CPU] UID:21783 Activity:high |
7/12 Is anyone running SETI@home on IRIX? I tried running the
setiathome-3.04.mips4-sgi-irix6.2 version on an IRIX 6.5.4m machine
with an R4600 processor, and it complains "Program not supported by
architecture." What's wrong? Thx.
\_ You already asked this question. It's self-explanatory. It's
apparently a MIPS4 (64-bit) executable and your processor is
a 32-bit chip. Whoever compiled it should've used the -o32
option to generate an (old) 32-bit object. --dim |
| 2001/7/9 [Computer/HW/CPU] UID:21742 Activity:nil |
7/8 I guess Intel reall doesn't want to sell any 64bit chips:
http://www.theregister.co.uk/content/3/20215.html
\_ Makes Sun's prices look almost reasonable. |
| 2001/7/6-7 [Computer/HW/CPU] UID:21723 Activity:moderate |
7/6 pentium 4's are dirt cheap now. should i get one? how are they better
than p3's?
\_ Butt memory expensive.
\_ no. they are broken wrt multiplications. - paolo
\_ URL please?
\_ href:mailto=demmel@cs.berkeley.edu
\_ Why use a wimpy old URL when you can talk to Demmel
himself?
\_ so is it better to get a P3 1 GHz or pay $100 more for a P4 1.3 GHz?
\_ No, buy an Athlon 1.3Ghz for < P3 1Ghz. Using a P4 has two major
problems: #1 The P4 has a piddlingly small cache. This will
require a major design revision, and probably won't happen until
the next process shrink. #2 One word: RAMBUS.
Oh, yeah, while I'm at it, the current P3 design will NEVER go
faster than 1Ghz. Intel already has the next revision ready,
and it will run faster than 1Ghz, but the current speculation is
that they won't release it for a while because it will tank the
already mediocre P4 sales.
\_ P4 isn't MP compatible either AFAIK. Go with AMD. Athlon
is much faster than P4 and there a reasonably priced MP
\_ Yeah those are P4 Xeons. ($$$)
boards for them as well.
\_ Yeah those are P4 Xeons. ($$$) Don't you have to buy
special Athlon multiprocessor CPUs also?
\_ Right now you have to buy a Athlon MP part, but
its not that expensive (~ $100 for 1.4 GHz). Its
supposed to get cheaper in August since there are
new Athlon MP parts and mbs coming out. I'm waiting
till then before deciding on what system to buy.
\_ If you want to buy a P4 the 1.4 GHz part is the sweet spot
at $195 retail. This compares to $165 for a retail 866 MHz P3.
\_ http://gamespot.com/gshw/stories/flat/0,12880,2781969,00.html
Notice how the Athlon 1.4 GHz stands up.
\_ Notice that the P4 is a fledgling product with a lot of room
for growth. Buy a P3 1.0 GHz now or wait until the P4 grows
up. Don't bother with AMD. --dim
\_ This applies to all AMD CPUs.
\_ The AMD Elan series are pretty good for their prices.
\_ The P4 is an abomination. Avoid it until it gets fixed.
\_ By the way, I have a GenuineIntel Family 6 Model 8 Stepping 3
800MHz. Is it a Pentium II, III or 4?
\_ 3.
\_ Hmm, I have a 6-5-1 which is a P-II. So Intel didn't up the
family number when it went from P-II to P-III? |
| 2001/6/30-7/1 [Computer/HW/CPU, Computer/HW/Drives] UID:21687 Activity:high |
6/30 what's your opinion on seagate hard drives? also,
athlon vs. duron, besides the caching issue, what
other differences are there in terms of performance
and quality? finally, do you know of any motherboards
that you'd recommend?
\_ I had bad luck with barracudas and cheethas, other
models seem okay. I don't know about the x86 stuff,
I generally avoid it. ----ranga
\_ If it's for a desktop machine, then the criteria I'd use to
select a hard drive is it's noise level. Seagate's new
Barracuda IV is the new champ (as soon as you can actually
*buy* one, that is) at 2.0 to 2.4 bels. If you can't wait,
then Samsung makes some that are in 3.0 to 3.1 range.
If you're into overclocking to improve your frame rate in
whatever first person shooter everyone is playing this
week, then you probably want to ignore me and get something
that spins at 14.4kRPM.
\_ Check Tom's hardware for information about motherboards.
The Asus A7M is pretty solid. It's based on AMD's 760
series chipset. Don't confuse it with their similar, albeit
inferior model based on the flaky AliMagick chipset.
Seagate disks aren't what they used to be. IBM disks tend
to be very reliable.
\_ Also look at the Abit KG7-RAID (ignore the RAID part), which
has the same "north" chipset (AMD 761), but can hold a whopping
(these days) 4 DDR DIMMs.
\_ Who now owns Seagate (as in the company Seagate?)
\_ a group of private investors
\_ I thought Veritas bought them?
\_ they only bought out one storage sw division from seagate. |
| 2001/6/18-19 [Computer/SW/Compilers, Computer/HW/CPU] UID:21569 Activity:nil |
6/18 So GCC 3.0 is released. Anyone know where I can find detailed info on
the contents?
\_ it doesn't build for mips-elf
\_ I'm more interested in the detailed changes of the release
(improved std C++ etc.). And if I were to build it, it would be
on win32.
\_ http://gnu.org ?
http://www.gnu.org/software/gcc/gcc-3.0/features.html
\_ EVIL COMMIE LIBERALS. USE VISUAL C++. |
| 2001/5/29 [Computer/HW/CPU] UID:21380 Activity:nil |
5/30 AMD Dual mb shipped or soon shipping:
http://www.theregister.co.uk/content/2/19272.html |
| 2001/5/14-16 [Computer/HW/CPU] UID:21268 Activity:high |
5/14 AMD announces mobile chips. Yay for competition.
\_ Whatever. x86 is for girls and kids, Real Men
use RISC!
\_ What do women use then?
\_ women don't use computers, they are barefoot and pregnant
and in the kitchen fetching men beers.
\_ Where's *my* barefoot, pregnant, beerfetcher?
\_ you've got to be capable of getting a woman pregnant
first. I hate beer. Mine's in there cooking and
cleaning.
\_ hehe, i make mine work, so I can stay home and
chill all day. =)
\_ Which corner of what city? I've got $5 burning
pure RISC or CISC chips but cam the stupid
a hole in my pocket and want to do something
nice for a homeless guy.
\_ Get out there and make me some money, bitch!
\_ Isn't it true that there's no such thing as true RISC or true
CISC anymore, that they're both converging?
\_ Yes. This knowledge makes it safe to ignore the kids who
think they're kewl because they use RISC chips and not a
recent x86 or clone which is basically a RISC chip.
\_ You're an idiot. x86 chips are not RISC (yes, I know
RISC/CISC is a stupid gradient and that there are no
pure RISC or CISC chips but can the stupid
technicalities) they have a CISC ISA. In this case,
it's because their instructions are non-orthogonal
and of variable width. They also have non-symmetric
pipeline orders which makes it difficult to employ
modern microprocessor techniques to speed up processors.
As a result, most modern day x86 processors employ a
decode stage microcode expansion to make it execute
much like what a RISC processor would. The goal of
the designers is to design a RISC-like chip but history
prevents them from straying away from their original
CISC ISA so they now have to implement this kludge
which works but is not very ideal. By the way,
those of you who still think that RISC means "few
instructions" while CISC means "many instructions"
are morons who don't read the book or pay attention
in class.
\_ So what does RISC and CISC mean now ?
\_ don't forget the 8 general registers instead of
an yother architecture's 32+ registers.
\_ The fact that x86 clones dumb down the RISC core speaks
volumes. Real men use RISC not RISC with CISC on top.
\_ but even RISC uses RISC with *ISC on top
\_ It's nice that you can babble out a bunch of stuff about
RISC/CISC in an effort to sound smart, but you really
should have stopped at the part about the gradient
between the two and that lack of true CISC or RISC chips
today. Then you might have sounded smart. But, no, you
just had to continue on and counter your own best point.
Where was I? Oh yes, you're an idiot.
\_ Oooh, you're so cool because you obviously have no
clue and are one of those loud-mouth morons who
just like to insult other people without having the
balls or the merits to back it up. "I'm cool because
I read the CISC/RISC article posted on Slashdot. I'm
technically competent because I work at a b2bi
startup doing ASP programming". You're a moron.
You're fucking clueless because you obviously didn't
even attempt to refute a single thing I said.Ã…
\_ You refuted it yourself just fine. I pointed that
out already. Thanks for making it so easy.
\_ real men fab their own chips, and run them on potatoes.
\_ real men don't use terminals either, they whistle 300 baud
tones down the telephone line.
\_ feh! real men program by yelling series of ones and zeroes
into the cable.
\_ Real men are somewhere else and not blathering about baud
rates and the CiSC/RISC gradient on the motd.
\_ like me! .... oh wait... |
| 2001/4/18 [Computer/HW/Memory, Computer/HW/CPU] UID:21010 Activity:very high |
4/17 I'm between jobs and want to build a fast Athlon Thunderbird box.
Any suggestions? I'm not sure about which motherboard to get
(KT133A?), what the diff is with the T-bird 200 and 266 MHz bus
versions, and whether PC133 or DDR memory is what I need. Thanks!
\_ What does being between jobs add to your query? Does that mean
you want one on the cheap? --dim
\_ stop being impertinent.
\_ Go to http://www.anandtech.com Lots of info there. They just did a review
of KT266 vs. all the other DDR 266 MBs out there. The forums are
of great help too.
\_ Thanks. The message I took home was:
"If anything, you should still stay away from the KT266, the
platform is simply not mature enough." So I guess I'm
going with a KT133A chipset, probably an Asus. After reading
a # of articles, I'm still not sure about whether a 200 or
266 is what I need, and whether the KT133A supports DDR.
\_ Um. There are othe choices than those two. Go with an
Asus A7M266; it's gotten good reviews, wins all of the
benchmarks, and isn't based on that chipset.
\_ What is "those two"? Thanks.
\_ Currently there are 3 chipsets:
AMD 760
Ali Magik something or another
Via KT266
\_ KT133A? KT133? Are these older?
\_ Yup.
\_ The two stated: KT133A (old tech) and the KT266
(unstable)
\_ Is there a winning combo for Athlons? For Intels, you
get a Pentium 3 866 with PC133 CAS2 memory on a CUSL2-C
and whatever video card you want. (since the Pentium 4's
blow, currently, and the higher-speed Pentium 3's have
heat problems, to say the least)
\_ just one word about the KT133A, dont get any orb-type fan with it
it'll bash and cut the capacitors situated right next to the CPU
\_ I think this fall will be a much better time to buy, fwiw. The past
six months have not brought enough advancement to be worthwhile.
Laptops will get more choices by then too, with amd and ati getting
new chips out.
\_ "later" is always a better time.
\_ You keep talking about this filling some need but never say what
your need is. Is this a game box, a cpu cruncher, a web server, or
just a random toy because you're bored? http://www.tomshardware.com is a
good place to start for opinion/info/benchmarks.
\_ Get the 266 FSB. This means that the cpu bus is running at 133MHz
(or DDR if you have DDR RAM), and the mult is much lower. The 133
bus wins all around over a 100mhz bus, imo. Also, it would behoove
you to get a system that can eat the $50 256M DIMMS. This memory
can only be used by certain chipsets. I have a 1000/266 TBird with
$512M of the cheap RAM. It's on an ASUS A7V133. --sowings
\_ jee-zuz. That will get you, what, 2500 terabytes of RAM?
\_ All of you giving advice don't have a basis for it. You don't know
what he wants it for. |
| 2001/4/7-8 [Computer/HW/CPU] UID:20892 Activity:very high |
4/7 I recently added a second processor to a win2k system. although the
motherboard recognizes it, win2k still says it's running on a
uni-processor system. do I need to do anything special to enable
SMP? (and please no "Yes, install *BSD" or "D00D, US3 GN00 L1NUX!"
answers...)
\_ You need to re-install Win2K. Win2K installs a different
kernel depending on how many processors you have at install time.
kernel depending on how many processors you have at install
time.
\_ You may have to also shell out more $s to Bi11 G4735 because
the unprofessional version doesn't support 2 procs (at least
my cd didn't, and no its not a w4r3z copy).
\_ Win2K Academic Version or something? Win2K Evaluation?
\_ he meant the "Professional" version isn't professional.
\_ Yes, install *BSD
\_ , US3 GNU L1NUX!
\_ 2k Advanced server. Domain "forrests."- paolo
\_ Yes, install *BSUXD
\_ , US3 R3D H4T GNU/Linux! [ spelling corrected ]
\_ DOOD! U N33D 2 US3 R3D H4T GN00/L1SUX! 17 RUL3Z!
[ spelling and grammar corrected for locale ]
\_ the correct answer is to go to the device manager, select the
\_ I don't see it there?
\_ it's not on the Driver tab? it worked for me -- original poster
\_ All I see is ACPI.
\_ mine said "ACPI Uniprocessor PC"
\_ Hmm, that's weird. Oh well, I guess I'll worry
about it when I get a second processor.
Computer > ... Uniprocessor PC entry, and click the Update Driver
button. |
| 2001/3/12-14 [Computer/HW/CPU] UID:20754 Activity:low |
3/11 Anyone know how async logic works (url to papers fine)?
http://www.theregister.co.uk/content/2/17356.html
\_ http://www.sun.com/research/async/index.html
\_ Thanks. Any others?
\_ Follow their related links page
\_ Lookup :
1. Asyn. lab at Utah (try that in Google)
2. Asyn VLSI group at CalTech
3. http://www.avlsi.com (spin off of #2)
4. http://www.csl.cornell.edu (spin off of #2)
\_ Actually, http://vlsi.cornell.edu
I remember there being an AMULET processor done in
in England. There is also a book on asyn design
from Royal Philips. You can find it at http://Amazon.com.
It , however, is a big rip-off of #2.
Mail me if you want more serious info. -tyf |
| 2001/3/9 [Computer/Companies/Apple, Computer/HW/CPU] UID:20736 Activity:nil |
3/9 And you thought that Apple was a retarded paranoid company when
it came to trademarks:
http://www.theregister.co.uk/content/3/17454.html |
| 2001/1/24 [Computer/HW/CPU] UID:20406 Activity:high |
1/23 Where can I get more info on Intel's Merced?
\_ Not to be flippant, but try http://intel.com?
\_ At least he didn't ask "How to get info on Merced?"
\_ http://www.intel.com |
| 2001/1/18-19 [Science/Electric, Computer/HW/CPU] UID:20366 Activity:nil |
1/18 Typically how much power does a desktop PC use? I'm not talking
about the max. rating written on the power supply.
\_ CPU + monitor, 150-200 watts or so. -tom
\_ How much w/o the monitor?
\_ Wow, I thought it was more like 50 watts or so. In that case
I'd better turn off all the idle machines here. |
| 2001/1/15-16 [Computer/HW/CPU] UID:20319 Activity:moderate |
1/15 How can I detect the CPU speed of a Tecra 8100 running Win2k?
\_ doesn't Norton still have a simple system profiler?
\_ i think you mean to say "estimate" or "determine" not "detect". -ali
\_ Ok, how do I estimate the speed if I don't have Norton?
My Computer-> System Properites shows it's a x86 Family 6 Model
8 Stepping 3, but have no idea what that means. I
basically just want to know if it's a 650, 700, 750, etc.
Shareware would be good.
\_ http://www.h-oda.com |
| 2001/1/11 [Science/Space, Computer/HW/CPU] UID:20294 Activity:very high |
1/9 Where can I find a product that will let me brainstorm in
the shower? Think a waterproof whiteboard, not a saran-wrapped
pPilot.
\_ How about a Pentium III with a 100W power supply and 20"
monitor. Those were water proof last I heard of it. And make
sure there's lots of salt.
\_ you have a one bit binary computer with you. do like most sodans,
and program that while in the shower.
\_ one-handed use product?
\_ quicker showers!
\_ Yeah. Conserve water, man!
\-fogged mirror --psb
\_ you're funny. and practical. listen to the man, people.
-psb #3 fan
\_ Truly. This would be so. Wisdom has come to you.
--psb #1 Fan
\_ I've become a fan of psb #3 fan.
\_ external (outside of the shower) tape recorder
\_ camcorder, even.
\_ webcam so we can teleconference
\_ etch a sketch! |
| 2000/12/17 [Science/GlobalWarming, Computer/HW/CPU] UID:20112 Activity:nil |
12/14 Just want to comment on Brodersen (cs152). Total bull shit.
Last lecture says "general purpose CPU is the WRONG way
to go. Parallel computing is the wrong way to go. Software is
the worst thing in the world. DSP s the way to go. By the way
BDTI and BWRC are hiring"
\_ For what? I took 152 with Kubi and he brought Brodersen in
as a guest lecturer. Broderson specializes in low energy
mobile devices so for that matter it makes sense to say
that a custom ASIC design that lacks any general purpose
characteristic is the proper way to go. By the way, what
does the W in BWRC stand for? That should answer your
question.
\_ who the fuck are you? nice degree.
\_ perhaps he's sick of buggy software coming from you people.
\_ "... you people". RACIST!
\_ that'd be SPECIEIST!
\_ No. Clearly that's RACIST! As in "You Asians all
look a like". Or "You blacks are all crackheads".
See? RACIST! Races are not species.
\_ uh, who's the crackhead here?
\_ I bet he meant that there will be diminishing return on investing
in uni-processor system. Parallel, distributed computing
would make greater use of existing hardware. (which is
a much stronger statement than is said in Patterson's book..) |
| 2000/11/20-26 [Computer/HW/CPU, Computer/SW/Editors/Vi] UID:19855 Activity:nil 63%like:19789 |
11/16 Thanks to AMD for donating Soda Mark VI, an Athlon/700 with 768M RAM.
Thanks to the CS dep't & the TDA project for the 160+GB of disk. |
| 2000/11/17-19 [Computer/HW/CPU] UID:19823 Activity:kinda low |
11/16 Who donated to the Mark VI project? They can't all be donated by
AMD and the Tertiary Disk Project can they? Also what happened to
the old parts?
\_ hey, now we'll be free from the selfrighteous "Alumni bought you a
shiny new soda" bullshit
\_ actually a complete system by AMD
\_ Yes. AMD gave us a system with 700MHz CPU, 18G disk, 768M DRAM.
The TDA gave us "a lot" of 9G disks and boxes for them.
Mark V is sitting around somewhere, essentially idle. Everyone
wants to take a break from it for now, failure-ridden POS
that it is at the moment. --PeterM
\_ If it's such a POS then why not make it an nfs'd /tmp
directory or use it to store mail for mail spool hozers.
\_ Do I hear a volunteer????? Cool. It's likely we'll
need a new VP next semester. Please feel free to run!
--PeterM
\_ Where can I get old parts? I'm building a tiny system for my mom.
\_ I got a large system for yermom right here. |
| 2000/11/17-19 [Computer/HW/CPU] UID:19815 Activity:high |
11/17 Can anyone give a simple comparison between the PII, PIII,
K6-2, K6-3, K7, Athlon, Duron, Celeron?
Athlon = K7 > PIII > Duron > Celeron > K6-2 > K6-3 > PII
\_ Why is K6-2 > K6-3?
\_ It's not.
\_ Athlon = K7 > PIII > Duron > Celeron > K6-2 > K6-3 > PII
\_ Why is K6-2 > K6-3?
\_ It's not.
\_ isn't PII > Celeron? I thought Celeron was a PII without
a special cache. Yeah, and I think K6-3 > K6-2
\_ Yep, and PII > K6-{2,3}
\_ new try:
Athlon=K7 > PIII > Duron > PII > K6-3 > Celeron > K6-2
\_ That's a big difference dumbass.
\_ fyi, the two tries are from two different
people. |
| 2000/11/16-17 [Computer/HW/CPU, Computer/SW/OS/FreeBSD] UID:19794 Activity:insanely high |
11/16 the new soda rocks! good job to all who contributed.
\_ Cool! So what's the difference btwn old & new?
\_ OLD: http://www.CSUA.Berkeley.EDU/computing/hardware/soda-mark-v.html
NEW: http://www.CSUA.Berkeley.EDU/computing/hardware/soda-mark-vi.html
\_ The linke to AMD on the new page is broken.
\_ Fixed, thanks.
\_ Do we still have the lottery scheduler?
\_ I think not. It didn't get ported to new kernel. --PeterM
\_ How hard would that be? I have some
free time coming up.
\_ I don't think it's necessary. Soda seems to be doing fine
without the lottery scheduler.
\_ Is there a BSD equivalent to "psrinfo" and/or mpstat?
\_ Have a look at the CPU and memory lines in /var/run/dmesg.boot.
\_ Do we still have to worry about /var/mail hitting capacity
as frequently?
/dev/da3s1e 8617749 1104247 6824083 14% /var/mail
\_ you decide.
\_ ummm, yes?
\_ Let's vote on it!
\_ Only if stupid people get to vote 5 times.
\_ With about 2500 users and 10 GB, people should still
limit themselves to 4MB.
\_ how about a /usr/contrib setup?
\_ Very well done. Thanks a lot for the time and effort, guys. -John |
| 2000/11/16-19 [Computer/HW/CPU, Computer/SW/Editors/Vi] UID:19789 Activity:nil 63%like:19855 |
11/16 Thanks to AMD for donating Soda Mark VI, an Athlon/700 with
768M memory. |
| 2000/11/1 [Computer/HW/Memory, Computer/HW/CPU] UID:19624 Activity:low |
10/31 When one CPU on a dual Pentium II machine executes an instruction that
writes to memory, and later the other CPU executes an instruction that
reads the same memory, how does the second CPU get the correct value
from the first CPU's cache? Thanks.
\_ Many cheepo/low end systems use a cache coherent protocol called
MESI (modify, exclusive, shared, invalid). This works fine under
smaller systems that can do bus snooping. But for much larger
systems (like 512 processor systems) they use a more sophisticated
directory lookup structure. -jeff
\_ For more information about MESI on intel processors, see
http://developer.intel.com/design/pentiumii/manuals/24319202.pdf
chapter 9, especially section 9.4.
\_ Distinction between cache and system memory. Dirty|clean bits.
\_ Do you mean the two CPUs share the same caches instead of having
separate caches?
\_ No, each Pentium has its own cache.
\_ They probably run in separate processes, separate
address spaces.
\_ You're probably a total fucking idiot.
\_ I believe that Hennessey and Patterson (not to be confused with
Patterson and Hennessey) discusses multiprocessor machines. You
you might want to check it out for a detailed explantion. -dans
\_ Just an aside, that naming convention is only used in Berkeley.
- current grad student somewhere else
\_ Uh, that is the author order on the books themselves.
Although I think Culler's parallel architectures book
does a better job discussing cache coherance. -nweaver
\_ Thanks for all the answers. I had a lousy prof for 152 (last name
started with R and was very long, don't remember exactly), and it
has been 7 years. |
| 2000/10/8-9 [Computer/HW/CPU] UID:19435 Activity:nil |
10/7 http://news.cnet.com/news/0-1006-200-2949585.html?tag=st.ne.1002.thed.ni |
| 2000/9/30-10/1 [Consumer/CellPhone, Computer/HW/CPU] UID:19379 Activity:high |
9/30 AAPL down 50% loss in market value 9 Billion
ATT donw 50% 110B
WCOM down 50% 87B
Sprint down 60& 30 B
Intel down 40% 250B
MSFT down 50% 320B
Dell down 50% 80B
Lucent down 65% 180B
FATALITY. This was all very hush hush over the last coupla hours
yesterday... I wonder what monday will bring?
\_ liu kang wins!
yesterday... I wonder what monday will bring?
\_ short-term thinkers are long-term losers |
| 2000/9/28 [Computer/HW/CPU] UID:19352 Activity:nil |
9/28 http://www.cnn.com/2000/TECH/computing/09/27/photon.chips/index.html |
| 2000/9/27-28 [Computer/HW/CPU] UID:19349 Activity:insanely high |
9/26 Sun released USIII released today. It seems to kick major ass
(of intel/amd/mot/dec) Anyone have a system yet?
\_ You're deluded. It doesn't beat alpha 21264 and only barely bests
Intel. It only excels in multiprocessing. --PeterM
\_ You're deluded. It doesn't beat alpha 21264 and only barely
bests Intel. It only excels in multiprocessing. --PeterM
\_ I bet Sun does a better job of selling their crap than Compaq
ever will. Compaq still doesn't understand all the gems
they have from that purchase they made a while ago. Compaq
weenies only understand NT and PC servers.
\_ Intel kicks the ass of every processor we have benchmarked
it against except for the new HP-PA chips. That does not
include USIII, but it was so much faster for our (intensive
number crunching) application than an USII 450 as to be
laughable. USIII would have to be twice as fast or more to
beat Intel, and these are only Intel 750's. --dim
MAYBE BECAUSE THEY CAN'T?
\_ Sun is selling 900Mhz USIII's now. SPECmarks
match a 1Ghz Intel on int and toast intel on FP.
\_ urlP
\_ I'll believe it when I see it. I don't trust SPEC
except as a very rough guide. So SUN's new processor
matches a 6 month old Intel chip, huh? That's supposed
to be impressive for the prices they charge? --dim
\_ 6 month old Intel chip? There are about as many
1GHz PIIIs in circulation as USIIIs.
\_ I've never seen intel kick any ass. MHz per MHz Intel
needs more than double to match PowerPC and Sparc.
\_ Duh! So why don't they DOUBLE THE CLOCK SPEED to
what Intel can do and BEAT INTEL INTO THE DUST?
MAYBE BECAUSE THEY CAN'T? Duh! CPUs have to be
DESIGNED to clock fast! Duh! MHz is MEANINGLESS.
\_ Okay I should have said it this way, they need
more clock cycles to execute the same number of
instructions because while US/PPC can execute 4
int instr./cycle and 2 fp instr./cycle Intel can
only do 2 and 1 (I believe). Bascially CISC is
wasting your time by doing less in more time.
\_ But RISC takes multiple instructions
to do what CISC does in 1. Comparing
instruction counts is as much as waste
of time as comparing Mhz.
\_ A RISC chip can introduce greater instruction
parallelism than a CISC chip can thus time
per instruction execution becomes important.
\_ Yeah, and they can't clock their complex (4/2)
RISC CPU fast enough to match Intel's relatively
simple CISC (2/1) in performance. Why is that?
It is because they failed to design it to
clock fast. Intel's is a superior implementation.
\_ No RISC is better. The only way that Intel can
keep CISC in the game is to keep ramping up the
speed. They can't improve performance in any
other way because of thier horrible design. Now,
\_
You seem to think that ramping up the clock
speed is trivial. You are wrong. The
CPU has to be designed from the ground up
to be highly clockable.
The MORE WORK PER CLOCK TICK, the longer
the logic stage, the less you can increase
the clock! It's a tradeoff. Designing
something to clock fast was a smart choice.
A winning choice. Intel designed
in such a way as to achieve high clocks.
Other vendors did NOT, or else you'd
see 1GHz PA-8500 uber alles. But the
fact is, no one can clock a PA-8400
at 1GHz precisely because it was never
designed to go that fast.
Only the 21264 is demonstrably "better" at
int stuff, and only barely. The PIII
and the K7 with their CISC are holding
up just fine.
RISC chips can get the same perf at lower speeds,
giving them room to expand perf in all aspects.
The high end RISC chips can run at similar speeds
as Intel's fastest but they can run thorugh more
instructions in the same time. The other problem
with x86 is the bad assumptions about uniform
memory fetch latencies. Intel has made some
pretty bad assumptions about how cache fetching
works and have spend a lot of time improving thier
pipeline, while the bottleneck is really the time
it takes to satisfy a memory fetch from either
l1,l2,mm,vm. It doesn't matter what your pipeline
is like if most of the time its empty.
\_ what a stupid measurement. That's like saying a particular
car is slow because it revs higher. -tom
\_ stop it Tom, you're making sense. :) -mtbb
\_ Retail cost should be a factor too
\_ I've had a SunBlade prototype on my desk since June (before it
was a SunBlade) - what do you want to know? -alan-
\_ Is the binary compatibily still there? There some problems
when moving to 64bit mode on USII (esp 200 MHz ones in the
UE2) and does it still run the 32 bit Solaris kernel (I
gather that 32bit kernel is faster than 64bit kernel).
\_ Yep. Complete binary compatibility (I have no idea
what problems you're referring to with the USII - it
was binary compatible as well.) No 32-bit kernel
though, 64-bit only. Which is faster depends on what
you're doing. -alan-
\_ For server work loads (ie non-compuational, non-db)
doesn't the amount of time required to search for a
given memory page in the 64bit address space cause a
noticable slowdown? And the problem with USII was
something to do with main memory fetch sometimes
failing on sub 200 MHz chip when the 64bit kernel
was running.
\_ Not that much of a slowdown, and on the plus side,
all 64-bit code is compiled with full UltraSparc
optimizations, while most 32-bit code is not.
\_ The only problem I heard of with < 200 Mhz Ultras
was the one where a bad instruction that no compiler
or assembler would ever generate could stall the
processer. That's a bug, not a compatibility
problem. (And fortunately for Sun, the actual
instruction was never publically leaked, unlike
the Pentium F00F.) -alan-
\_ Dont make old men run sprints or marathons |
| 2000/9/27-28 [Computer/HW/CPU, Computer/Networking] UID:19348 Activity:very high |
9/26 ARG! I bought a modem that I thought would work with linux
(The AOpen FM56-PM) and it DOESN'T. Please recommend a 56K
modem that works with linux. Thanks.
\_ D00D R U ST111 U51NG A M0D3M? U SH0U1D T12Y CA813 012 D51!
\_ Um.. buy a friggin external modem?
\_ You lack clue. Go get more clue.
/- This thread cracks me up. It's so funny reading shit about CPU
/ speed when 99% of the people cares only about network bandwidth.
/ How will USIII improve my p0rn d/l time? Get a fucking clue.
| \_ DOn't worry your pretty little eloi head. Us morlocks
| (1%) will keep things running. Oh, I'd like fries
| with that, too, please.
\_ Stephenson sucks at non-fiction
\_ contrary to popular belief on Soda,
NEAL STEPHENSON DID NOT WRITE
"THE TIME MACHINE" which is where "eloi" and
"morlocks" comme from. Go read it; it's
| you can use any old fucking CPU and disk because it
eTextified and less than 100 pages.
http://www.jyu.fi/~otto/finncon95/timem10.html
\_ He is referring to 'The command line' book
which did violence to Wells' invention.
Eloi rule the world, not Morlocks. Look around
you, coder peon.
\_ That's what the Eloi want you to think.
| \_ Some people actually write the code that allows you to
| download your p012N you 311T3 Hax0012. They care about
| things like memory fetch latency and instruction
| execution. In terms that you can understand. If the
| 311T3 p012N site is running on a USIII you can ask
| it to give you more files per second.
| \_ time for CPU to process URL requests: 2 us
| time for network to send p0rn mpegs: 1 hr
| time to enjoy porn: priceless
| Network bandwidth is priceless. For everything else
| you can use any old fucking CPU and dick because it
| doesn't really matter!
| \_ it takes a lot of CPU to keep 100 or 1000 Mb/s
| links full.
| |
| 2000/9/19 [Computer/HW/CPU] UID:19287 Activity:moderate |
9/18 [Rephrasing]
I want to build a CPU-intensive, 4-processor Pentium III machine.
It does not need to be fault-tolerant, mission critical, or have
RAID. It just needs to be a fast MPC box with one parallel job on it
at any given time. Suggestions? -jctwu
\_ Some of the big vendors are selling 4/6/8+ processor machines
so of course it is possible. Most of them are using Xeon chips,
though. The only thing you're going to need here is the
motherboard and you're set. What else is there to it? --dim
\_ BTW, soda was once a machine with 22 386 processors. Don't know if
it was PC architecture though.
\_ AFAIK, there are no motherboard chipsets that support quad PIII
configurations. You have to go with Xeon and this is fscing
expensive. But if you buy the box from Dell or SGI, at least they'll
build it right.
\_ what kind of job? might be better off with two dual-processors
and an appropriate interconnect.
\_ Photoshop? SETI@home?
\_ http://pcl.cs.ucla.edu/projects/parsec -jctwu
\_ it will run over MPI, so you could easily cluster some
dual-processors. with linux you could run bonded 100baseT
and get >180 Mb/s between them. if you need faster comm,
you will need gig-ethernet or myrinet, and the switches
and stuff start to cost as much as the 4-way or greater
inter SMP. shop carefully. the big SMPs may suffer from
serious bus contention if they're not good NUMA boards. |
| 2000/9/18 [Computer/HW/CPU] UID:19276 Activity:high |
9/18 I want to build a 4-processor Pentium III machine. I'll write
an FAQ afterwards. URL|adviceP, please let me know. -jctwu
\_ Is it even possible? Maybe you can do it with a PIII/Xeon? --PeterM
\_ Like with two mother-boards? or one that has 4 slots?
\_ Possible, yes. One motherboard, possibly daughter cards. I'm
doing the web research but if anyone has 1st-hand knowledge. -jctwu
\_ Some of the big vendors are selling 4/6/8+ processor machines
so of course it is possible. Most of them are using Xeon chips,
though. The only thing you're going to need here is the
motherboard and you're set. What else is there to it? --dim
\_ How many power supplies? How many hot-swappable drive bays?
There's a lot of stuff in vendor ads, and I'd like to reduce
extraneous parts. -jctwu |
| 2000/9/16-18 [Computer/HW/CPU] UID:19262 Activity:moderate |
9/15 I've got an extra Pentium 200 MMX lying around and I've
been toying with the idea of installing Plan 9. Anyone
tried it? Any +/- experiences?
\_ Plan 9 is very easy to install.
\_ How fast is it? Would it run well on a P200? And
what sort of stuff can I expect in terms of a GUI
and graphics?
\-plan9 would run on a p66. --psb
\_ how well?
\- at least one of the p9 insiders for a long time
ran it on a motorola 68k sun3. i think it was dmr but
that might not be right. --psb
\_ Sell me your Pentium 200 MMX processor. I will pay you
$35 plus shipping.
\_ Well its more than just the proc. I've got the MB,
ram, a couple of HD's and a extra power supply. I just
don't have a case, but I was going to pick up a cheap
on a WS or HSC. |
| 2000/9/8-10 [Computer/HW/CPU] UID:19212 Activity:high |
9/8 I'm trying to buy a wicked fast PC. i don't need more than 512 megs
RAM and don't need more than 20 gig disk. i don't need spiffy grfx
(some texture mapping would be nice though). what should i buy? what
kind of ram, what processor, what kind of bus chipset? thanks.
\_ /csua/tmp/newpc
\_ thanks, but i'm guessing you're either stupid or you can't read.
i'm looking for a "wicked fast PC" where cpu speed is key. i
don't need a '"cost-effective" for-home or for-business PC'
i need something that runs really fast. for example, a 733 mhz
machine you're pointing out is clearly suboptimal. thanks for
trying to help anywys.
\_ you are welcome, now shut the fuck up and go away.
\_ if you want faster than that, then you want a dual using
an 840 chipset and RDIMMs. This is bucks but you can do
it. Look at http://anandtech.com for this. -jctwu
\_ RDIMMs are not fast.
\_ Sure. But if you want a dual without too much
headache you must get RDIMMs with an 840. -jctwu
headache you must get RDIMMs with an 840. I'm not
implying that the 840 or RDIMMs are fast, even
though it looks that way. -jctwu
\_ Right. If you want dual CuMine Pentium III's,
you are forced to get an 840, which forces you to
get RDIMMs. -jctwu
\_ Hmmm. . . I'm guessing you're either stupid or just grossly
misinformed on what makes a machine fast. CPU speed is rarely
the bottleneck in machine speed. Unless you're doing
something that is very specifically CPU intensive (and, for
that matter, not memory intensive-- why the 512 meg max?),
that 733Mhz processor will do just fine.
\_ i'm neither. i need fast memory, and don't need more than 512
megs of it. it's not a bizarre question. you will notice that
i asked "what kind of ram, chipset" implying that i understand
that cpu clock rate is not the only issue. and yes, i want
fast disk as well. jusxt not more than 20 gigs. apparently,
you folks are confusing disk and memory size for disk and
memory speed.
\_ So what are you doing that you need this wicked fast PC
for? You've said you don't need buff graphics, so what
is all that speed going to? Office Apps (heh).
\_ array processing (number crunching).
\_ What are you going to do with it that doesn't need disk or tons of
memory or even good video, but a lot of CPU? Sounds like you
should be building an SMP box or thinking about a high-end alpha
or something. Are you looking for FP performance or integer? |
| 2000/9/7-8 [Computer/HW/CPU] UID:19190 Activity:low |
9/6 I was trying out the CPUID(0) instruction on PC's. Pentium Pro returns
"GenuineIntel", while AMD SC410 returns "AuthenticAMD". Anyone knows
what strings processors from other vendors return? Just curious.
\_ there are no other intel cpu makers that matter
\_ Cyrix? Do PowerPCs support that instruction? -John |
| 2000/8/16-17 [Computer/HW/CPU] UID:19008 Activity:moderate |
8/16 What exactly does VLIW try to depict? Does it mean that (1) all
instructions within an issue window are independent or (2) all
instructions within an execution window are independent. It seems
that EPIC which claims to be VLIW more resembles #1 while most
of the Phillips/TI DSP chips more closely resembles #2.
\_ this is a SysAdm Farm, wrong place to ask this kinda question
\_ what do you mean by "depict"? As I understand it, all instructions
issued simultaneously execute simultaneously. Why would you put
them in the same issue window if they had some kind of hazards
between them?
\_ Well, #2 is inclusive of number #1. What I meant by #1 was
that it was just the multiple instructions issued per cycle
that were independent and not all the intructions in the
pipeline. For example, IA64 only restricts all 3 instructions
in the 128 bit issue to be independent but that's just it.
TMS320C67 and the Trimedia 1300 requires that all outstanding
(non-committed) instructions be independent of all types of
hazards. Obviously #2 saves on a lot more hazard detection
hardware but makes it nearly impossible to schedule code.
But which is the actual definition of VLIW?
\_ Asking what the actual definition of VLIW is like asking
what the actual definition of RISC is. It depends on who
you ask. |
| 2000/8/10 [Computer/HW/CPU] UID:18946 Activity:moderate |
8/9 How is the load average related to the %cpu utilization? I have observ
on my machine a high load average despite low cpu utilization.
\_ load avg is a decayed avg of the number of runnable processes.
%cpu utilization is the fraction of busy time (* 100). you can
have a load of 1 and a %cpu of 100 if the load stays at 1 long
enough. the reverse is less likely. it means that you had a real
large number of procs become runnable for a short period of time.
-dpetrou
\_ What this sounds like to me is that the load average is based
on the number of runnable procs in priority que the scheduler
uses. The deeper the priority que the higher the load average,
right?
\_ Yes. Also note that load avg on an SMP machine is different
than a single cpu. A quad cpu machine with a load of X has
4x as many jobs queued as a single cpu box with load X. |
| 2000/8/4-7 [Computer/HW/CPU] UID:18890 Activity:nil |
8/4 What's a fair price for a dual processor (200 MHz) Ultra 2
with 256 MB ram? I'm asking around $2400, is it too high/low?
Checked on E-Bay but similar machines have sold for between
$1200 and $3000 so I'm trying to get a better idea.
\_ Fair market value is what someone is willing to pay. I'm not
being a smart ass. That's the real answer.
\_ Fair market value is $1200-$3000. I am being a smart ass, but
that IS the real answer. There is no Kelly Blue Book for
old model Sun hardware |
| 2000/7/29 [Computer/HW/CPU] UID:18812 Activity:moderate |
7/28 The Athlon architecture is often described as "having lots of
headroom" as compared to Pentium. But what does changing
architectures buy you in the Mhz race? Why will the P4 be able to
make 100 mhz improvements where the P3 allows 33Mhz?
\_ In general, newer architectures are pipelined more heavily,
so for a given process, the clock to clock delay is shorter,
so the chip can be run faster. -nweaver
\_ Bigger P ratings are cooler. http://www.intel.com/pratingsrkewl |
| 2000/7/27-29 [Computer/HW/CPU] UID:18794 Activity:moderate |
7/27 I'm looking to upgrade the CPU in my old linux box. Its got a P150
and an intel mb circa 96. Does anyone know where I might be able to
get a reasonably priced upgrade (> $200)? I looked at evergreen,
but I don't know if my system is supported. I'm also not interested
in buying a new box. I have plenty of more powerful systems, I just
want to upgrade this one.
\_ You want to know where to find an upgrade priced more than $200.
Hmm. Pretty demanding there... let's see...
\_ Someone keeps changing it to greater than. I'm looking for
something less than $200.
\_ http://pricewatch.com. Look up Pentium II 300 to give you some
ideas. |
| 2000/7/19-21 [Computer/HW/CPU] UID:18715 Activity:high |
7/19 Any reason to pay the high price for Rambus memory?
\_ read the man page.
\_ own stock?
\_ donated brain?
\_ too much money burning hole in your pocket? even Intel admits
that sdram-133 is faster in most cases.
\_ Buying RDIMM means you have an effectively RDIMM-only
board, like ones with the Intel 820 and 840 chipset.
Unfortunately, Intel has been promoting the 820/840 line
as their flagship chipset for the last 8 months.
\_ the dual-channel rambus boards (the 840?) are supposed to have
the most buff PC graphics bandwidth available now. this matters
if you use opengl for real data (not static game models).
\_ Games these days don't use static models. The reason your
apps that use opengl for "real data" are so slow is because
the guys who wrote them don't know how to make them go fast.
Thank you for playing. -blojo
\_ um, no. we're the ones who wrote them, and they stream
geometry over the bus as fast as it can go.
games use display lists that can be cached on-card. -karlcz
\_ Don't listen to blojo. He is all talk. |
| 2000/7/14 [Computer/HW/CPU] UID:18676 Activity:nil |
07/14 Hello, what type of cpu are in these newfangled i-appliances?
http://www.zdnet.com/zdtv/freshgear/products/story/0,3679,2440937,00.html |
| 2000/7/6-7 [Computer/HW/CPU] UID:18600 Activity:high |
7/6 Anyone knows of a good AT-form factor (not ATX) motherboard for
use with Pentium IIIs or Celerons? I want to upgrade an older
system (166MMX) and need to change the motherboard for a faster
CPU but my computer case is for AT type motherboards, but they
seem to be a dying breed. Where can I find a good one of those?
Asus P2B-B, I believe.
\_ Asus P2B-B, I believe.
\_ Anus P2B-B, I believe.
\_ So you're going to keep your 5 y/o video, disk, and other crap
and just replace the cpu/mb? And you'll get top notch performance?
No. You're wasting your money. Either up to a K6-2 300 or just
buy all new everything. I upgraded my p5-166 by copying my
personal data to a newly purchased p3 system and installing unix
on the p5. Amazing what a decent little server a p5 still makes.
\_ DO NOT get a K6-2 300. They're defective. I've attempted
4 different linux/BSD installs and they ALL crash. That
machine is still collecting dust. Fuck AMD.
\_ Yep. Must be the AMD. Couldn't possibly be any of the
other hardware or something you fucked up. Fuck AMD.
\_ Right. That's why AMD offered to replace them for
awhile, but then refused to exchange mine.
Obviously because other hardware was fucked.
\_ I had lots of problems with AMD and Cyrix. Save yourself
a headache now or later... go with Intel, linux is
developed on it. Also, watchout that the 3com ethernet
card is the exact one specified as supported. I once
got a 3com905b instead of a 3com905, and it wasnt
supported. Get an intel etherexpress, and save headaches.
-- have 5 linux machines, have intel stock
\_ I can upgrade other stuff when there is a need. Right now,
I think if I don't upgrade my motherboard, I may not be able
to find any AT motherboard later. Upgrading the cpu/mb to
say a celeron is cheap too. My 4mb diamond stealth vram is old,
but it's good enough if I just run unix and don't play games.
2 gigs of HD ain't bad. I also really like my solid steel
mini-tower case from pcpowercooling and ain't ready to
dump it yet. I am getting a Dell because I don't want to
build a new system from scratch anymore, but I want to keep
my old system useful for a few more years, hence the mb/cpu
upgrade. Thanks for your advice though.
\_ Point being that you're not going to get that much out of
it so you're wasting money. Graduate, avoid grad school,
get job, buy new hardware, be happy.
\_ Nah, with maybe $130, I can make it a very decent unix
box. Why spend $800-$900 for a new entry level PC? |
| 2000/7/6-7 [Computer/HW/CPU] UID:18598 Activity:nil |
7/6 So What's the word on the AMD Durons?
\_ Aren't they released already?
\_ Go to sharkeys or toms hardware for this sort of info. It's
where the rest of us are getting our info. |
| 2000/7/6-7 [Computer/HW/CPU] UID:18597 Activity:high |
7/6 Are there motherboards that will take either a K6-2 or an Athlon?
(so i can be cheap and get a k6-2 and upgrade later)
\_ no.
\_ The difference in price between an Abit socket 370 and the abit
slot A MB is less than 40-50 dollars, and the difference between
a 450mhz-k62 and athlon 650 is less than 50 dollars.
\_ If the extra $50 means so much to you, wait 3 or 4 weeks. Prices
will have dropped $50 by then and you won't have to be a wimp. Or
you could wait 3 or 4 weeks, buy the cheaper system and have $50
to take a girl out somewhere nice, once. Never know. Might get
lucky. I think that's a better option for you.
\_ If you are so miserly, forget the girl, go for the sure thing
and spend it on hardware.
\_ Miserly? $50 is already way too much to spend on a date
with a college chick. Nothing miserly about that.
\_ Oops, i forgot! At Berkeley, the girl pays half,
*especially* if it's to "somewhere nice".
\_ The difference in price between a 450mhz-k6-2 and an athalon 650
is $125 (and there abouts where I'm looking, where are you looking?)
\_ http://pricewatch.com |
| 2000/7/5-6 [Computer/HW/Memory, Computer/HW/CPU] UID:18586 Activity:moderate |
7/4 What is the deal with the cheaper (not slot 1) P3 say in a Dell 667.
How do they compare to the reg slot 1 type? Are they socket 7?
Are they more error prone/slower what? Why did intell go backwards
in their design.
\_ FC-PGA?
\_ It certainly makes low profile form factor machines easier
\_ Call Dell and ask.
\_ the flip chip (FC) packaging dissapates heat much better than
the slotted versions. it's cheaper just cause it's cheaper to make
a chip (corn cheaps!) than a chip on a board. Same performance,
though (except of course, some P3's have a 133MHz system bus, and
some have 100MHz system bus, depending on whether the suffix
is EB or E, resp.).
\_ Go for the EBs. If it's a non-slot one, it's most likely an
EB. The cache is half the size of slot one but it runs full
speed and is on-chip. Don't get the E when you can get the EB
for pretty much the same price.. an almost 33% increase in memory
performance..
\_ Actually, I would look for an "E", not "EB". The EB's are
already pretty much clocked as high as they can go and go for
the same price as the E's. The E's can be clocked to hell and
back...avg clocker can get about 850mhz out of 600mhz (EB ~ 700).
Check the overclocking pages on the net for more info. I found
http://www.overclockers.com to be a good resource among others. |
| 2000/6/29-7/1 [Computer/HW/CPU] UID:18568 Activity:very high |
6/29 A typical 32 bit RISC architecture has 32 bits/instruction. That
means each instruction is 32 bits, w/6 bits op code and the rest
is operands, const, etc. RISC binaries are usually bigger than
that of variable instruction set CISC (e.g. x86, Java byte code).
So does that mean a 64 bit machine with binaries compiled for its
native instruction set have a much bigger size than that of the
original 32 bit counterpart because the operands and const are now
much bigger?
\_ no. that just means that your register sizes can be up to
64-bits wide without artificial concatination like in the
MIPS I/II ISA. MIPS III/IV which are both 64-bit architectures
still use 32 bit instructions but their registers are larger
and are capable of using 40/44-bit addressing. CISC instructions
on average, are not considerably smaller than RISC because of
their variable lengths. Keep in mind that many CISC instructions
actually exceed 32-bits.
\_ so what happens when I have two operands that are 32
bits, and would overflow on a 32 bit machine but not on a
64 bit machine? Also, do 32 bit operands get sign
extend? Also, wouldn't you need a 64 bit instruction
to do the following more efficiently:
(word1<<32|word2)+(word3<<32|word4)?
\_ What does this have to do with swelling biological tissue?
\_ There are special intructions to load values into
the upper portion of the 64-bit registers. I don't
know what they are off the top of my head but
download the MIPS IV ISA manual and you can find
out. In terms of doing things efficiently, it is.
You have to look at things from a statistical
point of view. Each 64-bit value comprise of 4
16-bits so theoretically you would need 4 instructions
to form a single value. But for the common case,
most of those 16-bits are just zeros so it's not
as bad as you think. And if you think about it,
it's much more efficient to do something with 4
instructions than to do it with 2 instructions
that are double the size because in the latter case
you force all instructions to double in size which
isn't necessary.
\_ just to add to that, many program store constants
in a certain part of the binary so it's a matter
of issuing a lw.d instruction.
\_ You use registers instead of immediates, dumbass.
\_ ali, is that you?
\_ no. obviously, you don't understand my style.
i only talk shit at people who think they know
something, give out wrong answers, and become
irate when you suggest to them that they should
add some uncertainty in their answers when they
clearly don't know what they're talking about.
i think the above is exemplary of what a good
motd exchange should look like. -ali
\_Because without ali here to keep things straight,
the motd would be full of anonymous trolls and
wrong answers and oh my! We all owe a big
debt to ali for keeping the motd the pure source
of factual information that it is! Three
cheers for ali!!!
\_ the arrogant punk speaketh again!
- anti-ali association (AAA)
\_ Why do you have a problem with arrogance?
So the man is smarter than you and knows
it. I mean it's not like it takes much
to be smarter than you.
\_ Arrogance is a bad thing. Being smarter
is not. Being smarter is not the same
thing as arrogance. Try a dictionary.
--smarter than you
\_ Well said.
\_ Why is arrogance a bad thing? A
smarter person is better than a less
smart person and should treat the less
smart person accordingly.
\_ Arrogance is a Bad Thing because if
we all lived the way you'd have it,
life would suck for everyone. You
especially, I figure.
--smarter than you
\_ Why would it suck? I am prepared
to defer to my intellectual
superiors. This lame appeal
to Kantian ethics will get you
nowhere.
\_ Because Dalai Lama said so!
\_ moron. - the flaming goat
\_ Actually, in sparc, for example, the cpu did 32-bit loads
from memory. So you wanted both the instruction and the
data to be contained in one 32-bit word. If I'm remembering
my horrible sparc-asm encounter correctly.
\_ that dash shouldn't be there
Or maybe that was the 64-bitcpu mess
\_ much more common than loading a 64 bit constant is loading
an immediate address with a small number of bits (like, say 10),
and have that address point to a long constant. on an alpha,
you see a lot of:
lda $1,some_const
ldq $17,0($1)
where some_const is literal label which takes only a few bits
to embed in an instruction. -ali |
| 2000/6/21-24 [Computer/HW/CPU] UID:18510 Activity:high |
6/20 For the purposes of user apps, what would the approximate performance
equivalents be for Sun's SPARC{5,10,20,Ultra10} on the Intel scale,
assuming same specs on everything but CPU? (user apps is meant to
refer to non-realtime, non-graphics/DSP intensive, non-float intensive;
just, say, an office suite)
\_ you are a stupid idiot with poor critical thinking skills.
\_ http://www.spec.org/osg/cpu95/results/res99q4/cpu95-19991206-03938.html
As far as user apps, you're going to have a hard time finding
the equivalents of Office 97 or PhotoShop on Solaris systems.
\_ Doh, staroffice, what a bloated piece of junk. My P166
runs MS Office apps faster all at once than a brand new
Ultra5 runs StarOffice.
\_ StarOffice or Applix, PhotoShop or Gimp, not that hard.
\_ don't start this holy war; that was meant as a hypothetical
situation. thanks for the spec95 link
\_ Most of the machines you list are 7-8 years old, around the time
\_ I realize that. It's not an issue of whether Sun or Intel
r00L3z more. These are the machines I have available
unfortunately, so I just wanted to see where they fit in.
\_ Ultra10 is Ok desktop box. SS5/10/20 are fine as long as
they have decent CPUs (+100Mhz), 128MB RAM, and decent
disks (2GB or larger).
\_ basically Intel is cheaper and faster CPU, but SPARC
workstations are used more often to do real work
because Sun offers a complete performance package.
\_ Would you call Ultra5 or Ultra10 a "performance package"?
of the 486 and very early (very hot) Pentiums (60-90Mhz).
\_ For me, one SETI@home work unit usually takes 26-27 hours on a
Sparc20 with a 150MHz "Ross RT626" CPU (whatever that is), while it
usually takes 11-12 hours on a PII 350MHz. Both are running text
versions. So maybe a PII 350MHz is more than twice as fast as a
Sparc20. But then this is floating-point intensive, so maybe it
doesn't count. --- yuen
\_ Did they do a lot of optimizing? I was doing 26.5 hours per
\_ I don't know, but on the PII it has been roughly the same
speed for me since v1.2. -- yuen
unit on a p2-450 6 months ago.
\_ Times were A LOT shorter if you disabled the graphics.
\_ Linux is 3-4 times faster at setiathome than
\_ Aw, it's just a fancy Qume terminal.
\_ Uh, i dont think that sunray is a workstation...do me
a favor: Type "/usr/platform/sun4u/sbin/prtdiag" and
"uptime" tell us the machine model, the amount of RAM,
the number of CPUs, and the amount of cache on each CPU,
the memory banks size and intrlv.
\_ caught by the troll, hook line & sinker.
\_ If anyone is curious: Sun Enterprise E4500
159 users, 10240 MB RAM, 8 CPUs(each with
8MB cache running at 400Mhz).
Never seen anything so fast.
\_ Except an E10000 that I had to myself for awhile :)
Windows 9x on identical hardware. -tom
\_ I was using NT. I don't think it had an option to
disable graphics. I got bored and uninstalled it
after they started feeding the same packets for
three days.
\_ A co-worker set up some seti stuff on my workstation.
It produced 5871 WU today. What does that mean?
I dont know what kind of CPU it has, but my computer is about
3/4th the size of a laptop and has "SunRay" written on it.
\_ You're using an old model SunRay - the new ones are much
faster and look just like a monitor or flat panel display.
SETI WU's fly through them. |
| 2000/6/12-14 [Computer/HW/CPU] UID:18447 Activity:high |
6/12 What's the difference between a 386DX and a 80386? I have a "386(TM)DX
Microprocessor Programmer's Reference Manual" and a "80386 Programmer's
Reference Manual", but they list different cycle counts for the same
instruction.
\_ It's the exact same thing. One of your manuals is wrong.
\_ 386DX is the "Deluxe" version of the 80386
\_ Hmm, so the "Deluxe" requires more cycles? "REP MOVS" is 8+4*CX
cycles on the 386DX but 5+4*CX on the 80386?
\_ No one ever accussed the Intel x86 designers of doing things
that made sense. Just look at the whole segmented memory
nightmare.
\_ 386SX is the "SuX" version of the 80386
\_ I vaguely remember one has the co-processor built-in and the
other doesn't. But then, my memory is pretty faulty these days.
\_ http://aol.pcwebopedia.com/TERM/I/Intel_microprocessors.html
BTW, it was the 486SX that didn't have a co-processor. 80386's
never had them.
\_ just don't mention anything like this when Kahan is around
\_ 486SX had an 8-bit bus, whereas the DX has a 16-bit bus.
\_ I'm sorry, but that was an incorrect answer.
\_ a 386DX is an 80386DX without the annoying 80 part.
\_ a 386DX is, in fact, a .25 micron copper-process 68030
\_ so what's a 386SX?
\_ .30 micron aluminum
\_ .44 magnum
\_ *giggle*
\_ Or maybe it is a .8 CMOS -eric
\_ the 386DX is the 80386, as god intended it to be. the DX
to distinguish it from the later, lamer 386SX which is a
386 crippled with a 16bit data bus (as opposed to 32 on the DX).
Both are essentially programatically identical. To confuse
matters further, there was a 486DX (normal) and 486SX but
the difference there was that the former had a built-in maths
coprocessor and the latter did not. Bus widths were the same.
\_ And there's a 486GX with 16-bit data bus.
\_ Then how come the manuals of 386DX and 80386 shows different
cycle counts for "REP MOVS"?
\_ Why do you care? No one uses 386's anymore and the timings
on modern CPUs are far different.
\_ Didn't the 486SX actually have the coprocessor on-chip but
was disabled by an extra pin?
\_ speaking of processors, what distinguishes a 0.18 and a
0.25 micron processor (such as when some articles discussing
the dreamcast processors chips vs. the upcoming ps2 processor
chips? the 0.18 is supposed to be better than the 0.25
but harder to manufacture...what makes the 0.18 better, what
does that number refer too? -- no-clue
\_ 0.18 and 0.25u refers to the minimum channel length of a
transistor. Obviously, the shorter the channel length,
the less resistance there exists in the pull up and pull
down network of a CMOS circuit. Changes in process
technology also effects other factors like the reduction
of gate oxide thickness which increases the gate capacitance
per area of transistor and, therefore, increases the gain
as a result of applying gate voltage. Also, among many
other things, the threshold voltage changes. Take 105
or 141 and they'll tell you all about it (but by all
means, avoid Neureuther). -jeff
\_ Width of 'wiring'. Smaller = less heat from lower
resistence but harder to make. Less heat = can run
faster before dying from heat problems.
\_ thanks, i didn't know what that number referred to. |
| 2000/6/4-6 [Computer/HW/CPU] UID:18410 Activity:high |
4/65 does anyone know where you can get the Sun Ultra-5 for ~ $800
for the student discount?
\_ I was mistakenly spreading rumors about this today, the student
deal is $1300 -jones
see: http://store.sun.com/docs/specials/education/education.jhtml
\_ maybe you can get it w/o monitor?
\_ No. It's always a package deal for that price.
\_ My original post never said it was just a "student discount".
Anyways, I think it's too late now. Just like with the free
online classes thru SunEd. But wait awhile, the promotions
always come back (in fact U-10s
\_ this sounds juicy. more info please?
\_ http://www.sun.com/edu is the main site for current edu discounts
\_ Someone did some sloppy motd editing and trucated this thread
\_ http://www.sun.com/edu is the main
site for current edu discounts |
| 2000/6/3-4 [Computer/HW/CPU, Computer/SW/OS/FreeBSD] UID:18395 Activity:moderate |
6/2 I've got a 300Mhz AMD K6-2 machine. I tried installing linux on
it, but that failed because of some known and no-longer-fixable
problem on the processor. Is there any reason to presume that a
BSD install will also fail?
\_ I've had FreeBSD running happily on an amd-k62-300 for a long
time, with no problems at all. -meyers
\_ HEATHEN! LINUX! BIKE! ALL FREEBSD LUSERS ARE FAT AND UGLY AND
DON'T GET ANY EXERCISE! BIKE! LINUX! |
| 2000/5/10 [Computer/HW/CPU] UID:18223 Activity:nil |
5/10 If anyone has an old style powermac (serial ports, NOT usb),
approx 200 MHz or so, and is interested in selling, contact me,
-nweaver@csua |
| 2000/5/8-10 [Computer/HW/CPU] UID:18204 Activity:low |
5/8 On an x86 chip and in real mode, how does software distinguish between
IRQ5 and protection fault when an interrupt 13 occurs? Thx. -- yuen
\_ once in the ISR, poll the hardware device to see if it generated
the interrupt. sheesh, back in the day, we had ONE ISR, and we
had to check every device.
\_ Thanks. What if the protection fault happens in the IRQ5
handler? Do I just have to ignore this case? -- yuen |
| 2000/5/6-7 [Computer/HW/CPU] UID:18190 Activity:nil |
4/06 For those of you who think clock speeds are everything:
http://www.aceshardware.com/Spades/read.php?article_id=14
\_ This article is worthless. It's barely an article and it certainly
doesn't give an example or explain why clock speeds aren't
everything. |
| 2000/5/5-6 [Computer/HW/CPU] UID:18177 Activity:low |
4/35 What is the avg CPI rate of a 2-way superscalar processor (in-order
execution with branch prediction, caching, etc...)? I'm just looking
for some ballpark numbers.
\_ are you asking for number of issues per cycles or execution latency?
\_ issues per cycle.
\_ it's around 1 or 2 on average. cpus these days grab blocks
of instructions at a time. |
| 2000/4/25-26 [Computer/HW/CPU] UID:18107 Activity:high |
4/24 I'm having some computer troubles. I turn it on and for a second it
seems like it's booting up, but then it just stalls. The only thing I
did was attach a new fan. There are some more details and I don't
want to clutter the motd, so if you can help me out, mail me. Thanks.
-edhui
\_ Please include details. Put it in a /csua/tmp file is you want.
\_ Details in /tmp/edhui
\_> cat /tmp/edhui
> cat: /tmp/edhui: No such file or directory
\_ Details in /csua/tmp/edhui
\_ Your CPU is fried. K6-2's are notorious for being super-hot
chips (cf. Unreal). Next time, use thermal compound. Just a fan
\- GAME OVER, DUDE --psb
is never enough for that CPU.
\_ OK, then most likely I'll be replacing the CPU. What
exactly is thermal compound? -edhui
\_ Pasty stuff that sig. improves thermal conduction between
the heat sink of your fan and the surface of the CPU
die. At times it means the difference between
65 and 30 deg C die-surface temperature. Buy it
at Fry's.
\_ It should have come with your chip/fan. It probably
did and you threw it out.
\_ There's a teeny tiny itty bitty chance that you simply munged up
a cable connection when installing the fan. Pull the fan, check
all cables. Try again when CPU is cold. Do not ice. If it still
fails, then yeah, like the above says, your CPU is probably shot. |
| 2000/4/24-26 [Computer/HW/CPU] UID:18101 Activity:high |
4/24 http://www.techweb.com/wire/news/1997/10/1011merced.html \_ whatever happened to Merced? \_ Intel feared that the UC Regents would sue them for using this name which was reserved for the next UC Campus, UC Merced, so they renamed the chip to "Itanium". \_ Is this for real? (Having been away from campus for too long.) \_ Please. Merced is the name of a river as well as a city. It can't be reserved for UC. \_ Then is a "UC Merced" for real? \_ Yes, just search the UC web site for the details. \_ Merced and it's successors have been dubbed "Itanium"--and is shipping in engineering samples. Expect to see it comercially \_ this is a giant corp, though. Much like a student group, it makes progress on a purely geological timescale. some time this year. \_ Along with Iridium, they are planning to let it fall out of orbit and incinerate later this year. \_ "later this year" is what they said 2 years ago \_ this is a giant corp, though. Much like a student group, it makes progress on a purely geological timescale. \_ Um, no. Two years ago, folks expected Merced *design* to be finished in the same year (though no one at Intel thought that). Design finished last year, followed by actual silicon, followed by engineering samples. There are actual chips out now, as opposed to feel-good statements about when things will be done. \_ Unfortunately for Intel, rumor has it that the chip kind sucks and missed the boat as technology chip kinda sucks and missed the boat as technology marched on. We'll see when it gets here. \_ And of course, while Linux, IBM, HP, SCO & Sun will have Itanium-ready 64-bit OS'es this year, MS will make you wait for Windows 2002. \_ Oh ya? I thought they were already doing builds on it? \_ Well, Intel learned a lot on the Merced probject, and so though the first rev might not be terribly special, it is being used to finish the OSes on it. It's successor is being done faster because of that learning. Expect the successor to come out shortly (< 1.5-2 yeras) after Merced. |
| 2000/4/24-25 [Computer/HW/CPU] UID:18100 Activity:moderate |
4/23 From http://www.theregister.co.uk/000423-000001.html : "The third factor is that AMD itself was restricting supplies of some of its microprocessors as they reached an end-of-line situation, in advance of its simultaneous roll out of its up and coming Thunderbird and Spitefire chips in June." So, I was wrong. It won't be a month before the Thunderbird comes out. It might be 5 weeks. Wow, I feel so lame and uninformed now, since I was off by such a stunning time frame, assuming AMD doesn't launch it early to further undercut Intel's bottom line. As I said, buying a K7 now is stupid. Go dual p3 with slower chips and your system will be faster for any form of real work. \_ Saying you were off by a time frame is like using a point instead of a vector. \_ A time frame of a week? You're turned in papers later \_ A time frame of a week? You've turned in papers later than that. \_ The person is commenting on the grammar, not the time frame itself. But it doesn't look like you care too much for grammar. \_ I was off by a range of time. Thus a time frame. Thanks for the input. OTOH, I never saw anyone post the URL showing my original "one month to Thunderbird" was wrong despite saying I was totally off. I'm happier to be wrong by a few weeks than to be the no-URL loudmouth. As far as grammar goes, you should check your own before attacking others'. Glad you took the time to comment and express yourself. |
| 2000/4/23-24 [Computer/HW/CPU] UID:18086 Activity:insanely high |
4/22 Ultra 10 440Mhz UltraSparc-IIi, 256MB DRAM, 2MB L2 Cache, 9GB disk,
floppy,cdrom,Creator3D Graphics,keybd/mouse/etc. for $3200 brand new
from Sun. Thinking of buying one, for personal use. What do you think?
\_ I'm so pleased that you asked. This machine gets 18.1 SPECint95,
22.7 SPECfp95. For $3500, you could get a dual
Pentium-III/750E, which gets something like 36 SPECint95 per CPU,
256M of memory, an Nvidia GEFORCE 32M video card, a DVD drive,
27G of disk, and a decent 19" monitor.
check out http://langmuir.eecs.berkeley.edu/~peterm/BB.txt
--PeterM
\_ I thought that Windows98 does not support MP.
\_ We won't runn Windows on it. Windows 98 is so that I
can install Winblows on another machine. --PeterM
\_ But you're running winblows 98 on it. EEEUUUWWWWWWW.
\_ No, you're running an OS that has useful applications that
are fully compatible with the applications your clients,
customers and other business partners have. Nothing else
matters.
\_ 1 256MB DIMM instead of 2 128's.. really.. you're going to
want the free slots. Ditch the SuperMicro and PIII and get
a K7V and Athlon 850 for less. Ditch Win98. Everything
else looks perfect.
\_ AMD doesn't do SMP. We can get more CPU power in less
space by going dual. Had AMD gotten out a dual K7 chipset
which worked well, my config. would have used K7. --PeterM
\_ please sign your name so l33t sodans can belittle you.
\_ Don't get an Athlon. Sheesh. At least wait until the
'Thunderbird' chips are out in a month. Lamer. Stop
Giving out the generic 'kewl' advice on the motd when you
giving out the generic 'kewl' advice on the motd when you
obviously have no idea wtf you're talking about.
\_ The Thunderbirds won't be out in a month. What was that
about having no clue what you're talking about?
\_ Athlons are out of stock most places now anyway.
\_ That was the last announced timeline. You know
better than post the URL or fuck off. Either way
it doesn't matter. A K7 is stupid choice right
now.
\_ Because it's faster and cheaper?
\_ No SMP. Loses ground at higher speeds due to
1/3rd speed cache. !/4th speed at 1ghz. And
because it has compatibility problems that
Intel's chips don't have. And because a
dual P3 will be cheaper *and* faster *and*
run all your software without compatibility
fears with a modern, non-beta motherboard.
How's that?
\_ Wait until summer when Sun releases the UltraSparc III workstations
and slashes prices on the older Ultras.
\_ Good. Another question, will SUN stop using those crappy video
chipsets with 4MB of ram and crappy IDE controllers with no
UDMA support both of which are obsolete even by PC laptop
standards?
\_ To be fair, the SuperMicro MB doesn't support UDMA or
4X AGP. Intel is lagging on the chipsets. The BX is still
a high-perf. Intel chipset though. --PeterM |
| 2000/4/5 [Computer/HW/CPU] UID:17923 Activity:moderate |
4/5 Previously motdites said "wait a few months to get an Athlon."
Been too busy to keep up with the joneses. Why? THanks!!
\_ The chipsets on the motherboards sucked. A good one is out now
(Via KX133). To pick a good example, the new Asus K7V motherboard
(NOT the K7M) uses this chipset to good effect. --dbushong
\_ Speaking of which, when is the next Athlon rev. expected?
Worth waiting for? What about the "low end" Athlon?
\_ Got a URL? I was thinking of getting an Athlon and was told
the same thing--I don't have much PC hardware fu. -John |
| 2000/3/22-23 [Computer/HW/CPU] UID:17821 Activity:very high |
3/21 I've always built Intel but now I want Athlon. Athlons and their
motherboards are reliable right? What specific CPU / motherboard
would you recommend? Pointer to performance comparison so people
don't bonk me on the head would also be appreciated. Thanks.
\_ No. Wait 2 months.
\_ Agreed. At the very least, do *not* buy a MB based on AMD's
Irongate chipset. Get one based on VIA's KX133 (such as the
soon-to-be-out Asus K7V)
\_ Any reason you want to buy Athlon? -emarkp (Intel employee)
\_ [Ignoring the drivel below]
The Athlon is cheaper at the higher clock frequencies.
\_ But the cache runs at 1/3rd speed. The next ones out will
low performance, kinda like a rice-boy who spends 15K
be faster, probably full speed. That's why I said wait 2
months. Also the MBs are kinda shitty.
\_ Most RISC CPUs have full or 1/2 speed caches, and 1 MB
his civic even if he used his NO2s.
and 2 MB caches are pretty standard, rather than exotic
and "revolutionary" as they are advertised for older
CISC crap. Also buying Athlon means you have to deal
with sitty PC BIOSes and IRQ's and such. Sparc and PPC
mb's have all the latest stuff like USB, FireWire and
PCI while also giving you the advantages of OpenFirmware.
Buying a PC is just like buying a cheap car and then
tricking it. Its a waste of $s since what you end up
with underperforms the equivalent "professional"
version. Further more, clock speed is not everything.
Both PPC and Sparc execute 2 or 4 instructions per
clock cycle, thus at lower speeds they are more than
competetive.
unlike the athlon?_/
clock cycle, thus at lower speeds they can outperform
the chips from Intel and AMD. Also RISC chips like
PPC generate less heat and use less power, so compared
to x86 or Athlon, its like buying a v8 that gives you
the gas milege of a v4.
\_ note how you can't drive a formula one on public
streets, also i have an athlon. just got it and
it has 2 usb in the back besides RISC can't
optimize for crap anyway.
\_ Uhm yeah. I'm specifically talking about the 1/3rd
Athlon cache speed on the 1 giga hz chip. Not your
generic, knows-nothing-about-Athlons description.
Anyay, what the hell good is your Sparc or PPC or
whatever if it doesn't run my software?
\_ It's to impress all his geek friends. Geeks don't
run apps. They just like specs.
\_ Its more than specs. A Lose* machine
couldn't handle the kind of things I
ask my systems to do. A Linux/*BSD x86
box might be able to, but by the time
I added a decent nic, scsi etc they
aren't all that cost effective. Also
x86 based systems are saddled with
the crappy pc bios and all of the other
problems associated with it. I recently
bought a laptop with USB, FireWire, VGA
out, S-Video out, dvd a 5 hour battery
life and wireless 11 Mbit ethernet (that
has encryption enabled). Guess what, it
isn't based on a x86 (or clone) chip. In
fact most x86 vendors can't even match
all these features.
\_ What things do you ask of it? The box
you spec'd out is nothing special. And
what's wrong with the pc bios? You
tried out a pc bios written in the last
15 years?
\_ Yes. I have a pc bios for my x86 box
that was written circa 1996. It is
in no way as easy to use as open
firmware on my mac and my sun. show-devs
.properties, printenv and a full forth
interpreter blows the crap out of a
any bioses I've every seen.
\_ That's nice. And uhm exactly how
much time do you spend in the bios
writing forth code? You're coming
off sounding like a geek which is
normally pretty hard to do on the
motd, given the audience. We're a
long way from 1996 and thank god I
don't have a stupid forth
interpreter in my bios. What next?
MS Basic? Get real.
an Athlon. PPC is much better.
\_ Because he doesn't know anything about processors and wants
to buy a CISC x86 clone instead of a modern RISC processor
like Sparc, PPC (RS6000,Power4) or Alpha. He's all about
low performance, probably a rice-boy who spends 15K
tricking his 12K civic so its a rice-rocket civic racer.
If he decides to put all is money together (27K) he could
by a real man's car with at least a v6 that would trounce
his civic even if he used his NO2s. -- driving American v8
and computing with Dual PPC's and Dual UltraSparc 2's.
\_ PPC's rock! Down with Intel!
\_ And how many miles per gallon do you get on your American v8?
\_ Roughly 23. Depends on how fast I drive same as anyone
else. Never below 20. Never above 25. Always have AC
on.
\_ Mine give 15 to 20. But I care about performance not
about fuel efficiency. If I wanted power/thermal
efficiency in a processor I wouldn't pick x86 or
an Athlon. PPC is much better and faster.
\_ Thermal efficiency is important for laptops. I
don't care if my cpu will blacken an egg so long
as it doesn't burn itself up. Heat the whole
damn building for all I care. Power, baby! Gimme
speed!
\_ Or spend the $27K on a Prelude from the same manuf.
\_ A Civic weighs a lot less once you strip it. A lot of
racers then swap engines for a Prelude or Acura
engine. It makes for a fun hobby, I guess. --dim
\_ RICE BOY ALERT!
\_ Nah. My brother is into all of that shit. --dim
\_ Slap him awake. Do him a favor.
\_ And if the apps aren't available for the OSes that run on
those nifty processors, what's the point?
\_ QUAKE3!!!@!@!!11 |
| 2000/3/22 [Computer/HW/CPU] UID:17819 Activity:high |
3/21 For those in the working world, do you care at all wanting to know
what projects are in the pipeline, especially those you may
be working on or those you are interested in?
\_ Uhm. What the hell are you talking about?
\_ Mine more high school English.
\_ Don't know, since I know about all projects that are in the
pipeline.
\_ I don't care at all about wasting my time trying to understand your
mangled sentence.
\_ for some reason, this all seems humorous |
| 2000/2/7-8 [Computer/HW/CPU, Computer/HW/Drives] UID:17449 Activity:moderate |
2/6 Okay, I've decided to roll my own x86 computer, what are the best
price-comparison or e-commerce URLs you all can recommend to me? I
see http://www.sharkyextreme.com/hardware/weekly_cpu
listed below. I know about cnet and mysimon. What else is good out
there? What ones do you use? Thank you in advance, most wonderous motd.
\_ Just grab a wad of cash and go to the next computer show...
\_ When I was in SF, I went to both Robert Austin and the one at
the Cow Palace. I was underwhelmed, to say the least. They
had inferior-quality components, their prices were pretty
ridiculous considering it was a computer show, the entry price
was downright silly (at least RA sends out coupons) and the
selection stank. I'd rather recommend Spectrum computers in
San Francisco, or some of the similar Asian fell-off-the-back-
of-a-truck places. If you know what you want, the service
tends to be at least decent, sometimes even friendly, they
honor their warranties, they have good components and don't try
to bait-and-switch you too often, and the prices are good. I've
bought a lot of parts at these places while in the US, and also
sent non-technical members of my family there to pick up PC
hardware; none of us has ever gotten a bad deal. -John
\_ http://Pricescan.com. Buy parts from out of state and avoid the nasty
sales tax.
\_ http://Pricescan.com. We buys partsss from out of state and avoid the
nassssty saleses tax.
\_ http://Pricescan.com. Buy from out of state to avoid sales tax.
\_ Don't use places like Sharkey's for hardware recs. They have some
biases towards particular vendors and also aren't necessarily using
the system for the same thing as you. They use Quake1/2/3 as a
benchmark program, so go figure. Unless you're a Q123 player, they
won't be too useful for you. |
| 2000/2/5-7 [Computer/HW/CPU, Politics] UID:17438 Activity:moderate 73%like:17436 |
2/5 Why does the AMD K6-3 450 cost almost 3x as much as the
AMD K6-3 400 mhz version? (I mean it, both are K6-3):
http://www.sharkyextreme.com/hardware/weekly_cpu
\_ 1. Because yields are lower on the 450
2a. Because the 400 is priced to "lose" money
2b. Because it maximizes profit (see "Intel Pentium 3")
\_ Who cares? It's dead. Dead, dead, dead.
\_ Who cares? It's dead. Dead, dead, dead. Dead, Jim, dead.
\_ What exactly is the difference between a K6-2 and a -3? -John
\_ unless you don't feel like paying $600 for a CPU.
\_ You can pay less and get more.
\_ You some sort of throw back? You'd have us all using
8088s and 6502s? Those are cheap and slow too. |
| 2000/2/5 [Computer/HW/CPU] UID:17436 Activity:very high 73%like:17438 |
2/5 Why does the AMD k6-3 450 cost almost 3x as much as the
400 mhz version?
\_ Because cutting edge morons will pay 3x to get a chip that's only
about 3% faster. "Dude, leik i g0t another .3 FPS in Quake4 !!1"
\_ get a fucking clue. the K6-3 450 is not cutting edge. -tom
\_ Sure it is. What's not cutting edge about it?
\_ you mean other than the fact that you can buy
chips which are twice as fast? -tom
\_ Because they can charge it. The die size on the K6-3 is
about twice that of the K6-2, so unless they can charge
much more than previous chips, they'll lose money on each
chip. Since they have much smaller yield at 450 than at
400 that means the 450's are effectively much more
expensive to manufacture. |
| 2000/2/3-4/27 [Computer/HW/CPU, Academia/Berkeley/CSUA] UID:17413 Activity:nil |
12/09 "got mips?" t-shirts! See dans or visit the CSUA office. $10 |
| 2000/1/19-20 [Computer/HW/CPU] UID:17270 Activity:very high |
1/19 The Transmeta Crusoe runs at 1 watt of power (20 milliwatts idle).
What is the range of power consumption for mobile processors out
there now?
\_ It seems that the whole world is trying to get onto
yet another AMD/Cyrix/etc 1
http://www.transmeta.com
\_ somthing changed, but I can get on the site now
\_ It seems that the whole world is trying to get onto
yet another AMD/Cyrix/etc 2
the motd
\_ motd poll: what do you think of the future of transmeta?
yet another AMD/Cyrix/etc 3
another StrongARM 1
bought by intel/amd 1
the next Intel! 0
bought by intel/amd 0
dont care 1
hotter and sexier than tom 2
hey what about Motorola ?
\_ hotter than tom? does tom run under 1W and 20mW idle?
\_ Indurain had a resting pulse rate of 28 -tom
hotter and sexier than yer mom 1
hotter and sexier than sky 1
\_ SHE CANNA RUN THIS HOT FOR MUCH LONGER, CAPT'N
fad 1
pure hype 1
most likely to break VA Linux's IPO record 1
\_ Is there anyone who really considers their own mom hot and
sexy?
\_ No, just *yer*mom. My mom is elderly and gray-haired, and
stays at home knitting lace doilies.
\_ Of course. Don't you read the "real stories" on
alt.sex.stories.incest?
\_ I believe that most modern chips run at 10 watt/cm^2 or somewhere
in that ballpark.
\_ The hot K6's used to run at 40 W at 100% utilization, the
.25 um processors about a quarter that, the mobile 650 MHz
Pentium 3 uses 21.5 W
\_ Note that this is the 3120 chip which is designed for handheld. I
think the StrongArm runs at about 1 watt too. The new Pentium III
with PowerStep runs at about 6 watts when on batteries, and 11 watts
when plugged into the wall. (http://www.news.com
\_ WOW! now THAT's (microsoft style) Innovation!
They just invented the non-user-controllable turbo-switch!
\_ You can control it. Plug in/Not plug in. See? Innovation.
\_ The idea is to give more power when you're using the laptop
connected to the wall, but conserve battery life when you're on
the road. This allows laptops that are used more like mobile
desktops to have performance closer to desktops.
\_ and if I WANT to burn battery life on the road?
\_ The G3 is around 5.7 W, i believe.
\_ The G3 kicks ass...fuck Intel -- swings
\_ Maybe when you're inside the Steve Jobs Reality Distortion
Field (tm). In the real world it's nothing special. |
| 2000/1/9-10 [Computer/HW/CPU, Computer/SW/OS/Windows] UID:17196 Activity:high |
1.8 Location: SoCal. Need good provider of line to webhosting server
recommendations? - kinney
\_ Dump NT and use a real operating system. -- ilyas
\_ sure, want to port crystalreports to unix?
\_ he's running or going to be running at the minimum,
athlon 500, 256 MB memory 20 Gb disk. He's middle mgmt,
and has money to burn.
\_ Nah, he's only have to dump NT if he was using an Athlon...
\_ ?? NT won't run on K7's?
\_ It's running fine on this athlon 500. - paolo
\_ No it _isn't_.. shut _up_, paolo.
\_ It's running fine on this single proc athlon 500. - paolo |
| 1999/12/30-2000/2/3 [Computer/HW/CPU] UID:17124 Activity:nil 68%like:17049 |
12/09 got mips? The CSUA has a new t-shirt design. Front has got mips?
Contact dans or drop by the office if interested. Price is $10 |
| 1999/12/18-21 [Computer/HW/CPU] UID:17060 Activity:nil |
12/18 Jello. I am trying to dual boot a athlon-500 w/ nt and fbsd but am
getting a bunch of bsods and SIGSEGVs (during gcc) o
when i install / buildworld (respectively) Anyone have any luck?
I already tried turning down the memory timing. - paolo
\_ You're overclocking? Sigh.
\_ it's really sad to see that the only people care about these
days are clock speeds. if you overclock your cpu you start
to get diminishing returns in exchange for having a hotter
cpu and higher risk of failure.
\_ No I'm not. I see no need to overclock. I would just buy
a 750. Not My Money(tm)
\_ If it's not your money, buy an Intel dual or quad MB and
don't worry about it. |
| 1999/12/9-30 [Computer/HW/CPU] UID:17049 Activity:nil 68%like:17124 |
12/09 got mips? The CSUA has a new t-shirt design. Front has got mips?
design and back says "U.C. Berkeley Computer Science Undergraduate
Association". Contact dans or drop by the office if interested.
Price is $10 (we can probably arrange to mail you one for a few bucks
extra) |
| 1999/12/7-9 [Computer/HW/CPU] UID:17018 Activity:very high |
12/6 Is it worth getting a portable MP3 player?
\_ I got the RCS Lyra as a gift. it's pretty simple to use. Comes
\_ I'm looking to sell my 32mb Rio PMP300. I don't know what
it's worth, so make an offer. -jkwan
\_ I'll Buy That For A Dollar!
\_ I got the RCA Lyra as a gift. it's pretty simple to use. Comes
with 32MB card (64 MB cards are available). a cable connects to
both your parallel & keyboard port (not OR, it's AND), which i
found kinda odd. it uses RealJukebox to transfer the music, which
is quite fast.
\_ it connects to the keyboard port only for the higher power
capacity line available there. The communications are
all over the parallel.
\_ If you like the music that is published only in mp3 form, then
it seems like it's worth it. It would be a bitch to have to
rip and encode from CDs to get your collection.
\_ I guess its not worth it if most of your music is in
CD format. I guess CDR's are better if most
of your collection is in CD format.
\_ Huh? CD ripping is trivial. It's read time plus setup time
which is no different than CD-r setup time. Where do you
people get this stuff from?
\_ So what are people using nowadays (say on the windows
platform) to rip and convert quickly? Is there any
share/freeware program that does this to 128KBps
nicely/simply?
\_ Audiocatalyst is good for Windows
\_ is perfectly reasonable and legal to rip your OWN CD's for your
own use in things like a portable mp3 player. Its called
'fair use'.
12/7(?) Can anyone point me to a good CISC vs. RISC article / paper? Thanx.
\_ No, most designers avoid CISC if at all possible. Intel chips
allow for their legacy CISC ISA to run on their modern
chips but that's just a hack (CISC instructions are decoded
into micro-RISC ops). Basically, many things will not work
unless what you execute are RISC instructions (pipelining,
Tomasulo dynamic execution, etc...). Intel chips suffer in
that all instructions must pass through one extra level of
decoding which could effect branch prediction recovery and
because variable length instructions can cross cache block
boundries instruction causing miss penalties and rates to go up.
\_ Which is why the x86 line was predicted to die 10+ years ago
and oh wait, no, it's still here and still in 90+% of
computers from workstations on down and moving into the low
end server market.
\_ Wasn't the consensus that the argument is meaningless?
Most chips today have the best qualities of both.
\_ This is true. There are no more true CISC or RISC chips
being produced today.
\_ most chips now days have RISC-properties. people
today seem to be confused about what the terms
RISC and CISC really means so depending on what your
definition is, most chips are RISC.
\_ But with everyone adding extensions like UltraSparc
VIS & Intel MMX to the instruction sets, most chips
are CISC too.
\_ I'd say, "RISC-like". True RISC was only an ideal
and a theory anyway.
\_ Patterson & Hennessey
\_ something like "A case for RISC" from Computer Architecture
News" by David Patterson (1980). There's also a very famous
dissertation from a Berkeley student. If you give me some
time I'll look it up.
\_ - D. Patterson & D. Ditzel, "The Case for the Reduce
Instruction Set Computer," Computer Architecture
News 8,6 (Oct 15, 1980)
- Manoils Katevenis, "Reduced Instruction Set
Computer Architecture for VLSI," PhD Dissertation,
EECS, UC Berkeley, 1982.
--jeff
\_ "Why my x86 r00lz y0r powermac!!!11", Dudester69!11
\_ Here's your paper:
"Real World Consumers: by Consumer Man#232
A pc is $700. A mac is $1500. My 10 yr old kid can't tell the
difference, and the 15 year old tells me linux won't run on a
mac, oh and he wants a laptop, like in that movie he saw.
Gee, I wonder what I'm buying for christmas."
\_ PC R))LEZ!! MAC DR))LEZ!!!1
\_ Are what they said still more or less valid, now that
it's been two decades and the industry has been moving
fast?
\_ kubi seemed to have recommended the reading to us.
he didn't recommend anything more recent but probably
because more recent work in architecture isn't
focused on ISA as much as it was in the 80's
--jeff
\_ Even x86 implementations have some sort of RISC
core that is given a CISC interface. RISC introduced
follow-up letters at http://www.reason.com/9610/ltr.sl.html
a number of principles that are still considered
valid such as uniform instruction length, pipelining,
on and on.
\_ There's an article in a past issue of "Reason" magazine about
history's influence in choice between technologies. It didn't talk
about RISC vs. CISC, but it talked about PC/DOS/Windows vs. Mac,
QWERTY vs. Dvorak, Beta vs. VHS, etc. I don't have the issue with
me right now though. -- yuen
\_ Do you mean: http://www.reason.com/9606/Fe.QWERTY.html (and
follow-up letters at http://www.reason.com/9611/ltr.sl.html
[url fixed]
\_ _excellent_ article.
\_ Wow! Yeah, that's the article I read. I didn't know the
\- helpful. /- not helpful.
magazine has a web site. -- yuen
\_ Your tenses do not match. ^has^had
\_ actually that is just fine. He read the article in
past. HE didn't know the web site existed in the
present. Those are two different statements.
If you are gonna be a grammar cop get it right.
\_ Your tenseness is showing. Loosen up.
\_ Your value-add is null.
\_ Found this article on the subject:
http://www.ars-technica.com/cpu/4q99/risc-cisc/rvc-1.html |
| 1999/12/6-7 [Computer/HW/CPU] UID:17014 Activity:very high |
12/6 Is anyone speaking at the GM?
\_ Yes.
\_ 1'|| SP34|<, d00d! -B1FF
12/7(?) Can anyone point me to a good CISC vs. RISC article / paper? Thanx.
\_ Wasn't the consensus that the argument is meaningless?
Most chips today have the best qualities of both.
\_ No, most designers avoid CISC if at all possible. Intel chips
allow for their legacy CISC ISA to run on their modern
chips but that's just a hack (CISC instructions are decoded
into micro-RISC ops). Basically, many things will not work
unless what you execute are RISC instructions (pipelining,
Tomasulo dynamic execution, etc...). Intel chips suffer in
that all instructions must pass through one extra level of
decoding which could effect branch prediction recovery and
because variable length instructions can cross cache block
boundries instruction causing miss penalties and rates to go up.
\_ Patterson & Hennessey
\_ something like "A case for RISC" from Computer Architecture
News" by David Patterson (1980). There's also a very famous
dissertation from a Berkeley student. If you give me some
time I'll look it up.
\_ - D. Patterson & D. Ditzel, "The Case for the Reduce
Instruction Set Computer," Computer Architecture
News 8,6 (Oct 15, 1980)
- Manoils Katevenis, "Reduced Instruction Set
Computer Architecture for VLSI," PhD Dissertation,
EECS, UC Berkeley, 1982.
--jeff
\_ Are what they said still more or less valid, now that
it's been two decades and the industry has been moving
fast?
\_ kubi seemed to have recommended the reading to us.
he didn't recommend anything more recent but probably
because more recent work in architecture isn't
focused on ISA as much as it was in the 80's
--jeff
\_ "Why my x86 r00lz y0r powermac!!!11", Dudester69!11 |
| 1999/12/6 [Computer/HW/CPU] UID:17012 Activity:insanely high |
12/6 Is anyone speaking at the GM?
12/7(?) Can anyone point me to a good CISC vs. RISC article / paper? Thanx.
\_ Patterson & Hennessey
\_ something like "A case for RISC" from Computer Architecture
News" by David Patterson (1980). There's also a very famous
dissertation from a Berkeley student. If you give me some
time I'll look it up. |
| 1999/11/20-22 [Computer/HW/CPU] UID:16930 Activity:moderate |
11/19 Anyone have a link to a description of where the cache is on the
celeron vs. where it is on the pentium ][? A friend of mine is
trying to argue that the celeron's is not on the CPU, but is for the
pentium. That's just wrong, but I want definitive proof to fuck him up.
\_ I believe in early editions of the celeron, your friend is right.
However, in later editions of the celeron, they added a
cache. So perhaps you can both be right.
\_ perhaps this helps one of you:
http://developer.intel.com/design/celeron
\_ That's it: You're both right!
http://www.anandtech.com/html/articledisplay.cfm?document=277
Happy Days for Everyone! Now if only the Bears win the Big Game.
\_ not bloody likely. go trees!
\_ Well, L1 or L2 cache?
\_ The original Celeron processor had no L2 cache. It was essentially a
crippled Pentium II. The 300A version and later (originally
named Mendocino) has an integrated 128KB L2 cache on die.
The Pentium II processor has a 512KB L2 cache on the
backside, and runs at half processor speed (if chip is 400
MHz, cache speed is 200MHz). The Celeron processor cache
runs at full speed since it's on die. The new Pentium IIIs
(Coppermine) have a front-side cache of 256KB (I'm pretty
sure). -emarkp (Intel employee)
\_ Can you get me Intel chips for less than wholesale?
\_ yes. as long as you don't need them to be working.
\_ No fan is ok. Working is prefered. You can't even get
older stuff for cheap (and working)? |
| 1999/11/8-9 [Computer/HW/CPU] UID:16845 Activity:very high |
11.8 pop quiz - 83 people. $3000 up front, $10/mo or so, how do we get
fast net?
\_ But where?
\_ <DEAD>www.bandwidthcoop.com<DEAD>
\_ Do you live in the USCA Co-ops? Make it a capital expense off your
house's long term improvements budget and run as much net as you
can get away with which should "include future bandwidth growth
needs and scalability as high bandwidth content becomes more
common and even required on the net". You can get the Central
Level to eat %100 of the up front cost and probably most or even
all of the monthly charges if you do it right. -USCA Veteran
\_ the USCA? Ethernet jacks in every room? "It can't be done."
(or prove me wrong)
\_ weren't they gonna do this in cloyne, what, 4 years ago? 5?
\_ It can be done. Go read the alumni newsletters on their site.
(PDF). Gordon Moore of Intel is a USCA Alum and Intel is
donating some unspecified design and architecture skills to
the project. Wouldn't be surprised if they threw some hardware
at it as well. You'll be proven wrong. Probably not this year
as the USCA is *s-l-o-w* but probably in the next 3 or 4 years.
\_ unless the maintenance manager sells all the stuff to buy
speed. given the size of your co-op i think i know which
one youre in and i think you know what i'm talking about.
you can buy alot of 8-balls for the price of a p-III.
\_ The capital improvements budget is drug free. -USCA Vet.
\_ "Super."
(As you can see, I'm expressing the same wary hopefulness
I put into the Cal football team.)
\_ Just one more way to fluff workshift hours as the Computer
Manager and the Computer Manager's two five hour/week
assistants. No experience required, of course. A shared
network and a sniffer might prove interesting in most of
the houses.
\_ Why when switches are so cheap nowadays? --dim
\_ Because one hub per floor or per building or
whatever would add so much more grease to the
gossip mill. Not like everyone doesn't already
know who cheats on who with who but it could add
a whole new dimension to it. |
| 1999/11/6-7 [Computer/HW/CPU] UID:16835 Activity:high |
11/05 Today's the day: http://www.dcd.uscourts.gov/notice.html \_ if i had two pistols i'd run around yelling "yippie kiya" and shootings round into the air. and shooting rounds into the air. \_ The decision made you richer? I doubt that. \_ Damn straight. doubt away, baby. \_ There are other reasons to rejoice besides personal wealth. You might see that one day. \_ /csua/tmp/findfact.* \_ "The Findings will be posted in two versions, PDF and WordPerfect." No MS Word. :-) -- yuen \_ Gov't and law use almost exclusively WordPerfect & PDF. \_ Because MSWord doesn't count words correctly by the way the courts want word counts done. WordPerfect does. \_ You can still read WordPerfect document with MS Word. Hahaha! \_ "hahaha!"? Uhm, what decade are you in? There's been reasonable cross word processor import/exporting for the better part of this decade and it was at least ok in the late 80's. Are you stoned or something? I don't get what you're hahaha'ing about. \_ http://dailynews.yahoo.com/h/ap/19991105/ts/microsoft.html Yeah! \_ anyone notice that MS is releasing _its_ statement in WP format -- which is the first _reference_ to WP I've seen on their site? smells cheesy...... =} [http://www.microsoft.com/presspass/trial/] \_ pls see yuen's entry above |
| 1999/10/27-29 [Computer/HW/CPU, Computer/HW/Drives] UID:16779 Activity:high |
10/27 I want to buy a LINUX box. I don't need (or want) windows.
Any suggestions on online companies which sell LINUX boxes?
I could just use gateway or dell, but they charge me for windows
and the last time my company ordered something from Dell it had
some hardware that wasn't well supported by LINUX. Thanks.
\_ Dell sells machines preloaded with Linux these days.
\_ http://www.varesearch.com perhaps a little more high-end than most dells
but comes functional w/ linux. except they give you the normal
stupid soundcard that only does half-duplex w/ linux drivers,
so mbone multiconferencing doesn't work.
\_ I checked out VA Research, but for their prices I might
as well order from CompuUSA which is cheaper.
\_ Last time I checked CompUSA, they were selling Compaqs,
PackardBells and HPs. I don't know about HPs, but
Compaq Presarios and PackardHells are utter pieces of
crap and have problems even running windows, good luck
installing Linux one of those.. Seriously, though, check
out Micron and Dell websites, they make good systems
usually or build your own ..
\_ If you're running a linux box why are you paying more for someone
to pre-assembly it in a configuration that you won't be entirely
happy with? It is *very* easy to assemble a pc from parts today.
You'll get *exactly* what you want and not pay a dime to M$.
\_ Ok. If it is that easy maybe I'll try it. Do you have
any suggestions on where to buy the various pieces and
how to put them together? For example, when you buy a
motherboard and a hard drive, does either one come with a
manual telling you how to get them to work together? I'd
be happy to RTFM if I knew how to get my hands on one, but
"man motherboard" doesn't seem to work =).
\_ The problem with building your own machine is troubleshooting
it. A computer builder can at least make sure all the parts
work before giving you the machine. I like to work with
Central Computer: they'll build a machine to order for
you, and you can ask 'em not to install Windoze. Another
plus is you can go bring stuff back in person, if anything
is defective. --PeterM
\_ The stuff all comes with manuals. Mostly it's in English.
Pretty much each thing only fits one place in one way. The
red line on the HD (and floppy) data cable is pin #1. Do
it in a place with lots of light. It's pretty hard to totally
fuck up, but it may take an afternoon on your first time out.
Unless you've got money to burn and your time is incredibly
precious (or you're buying in bulk for a company), the self-
assembly thing is always a good bet. Do some price checks on
the net for the hardware/pieces you're looking for. Buy stuff
like the CPU/MB last. By the time you figure out what you
want for the rest of the system, cpu prices will have dropped
again.
\_ You should try http://www.computergate.com They have a fairly
broad selection of components at reasonable prices.
Certainly, you would be able to build a computer there.
\_ The first time is hard for everyone. You should do your
research, but here's a list to start you out.
ABIT BE6 motherboard Celeron 466 PPGA + "slocket"
IBM 9.1 GB UDMA/66 IDE HD adapter + PPGA fan
Intel i740 video card Creative Labs OEM 48x CD-ROM
Yamaha YMF724 PCI sound Side-paneled ATX mini-mid
card case
3Com ISA 56K Sportser Linux Mandrake 6.1 or
Yamaha YMF724 PCI sound ATX mini-mid case with side
card panels
3Com ISA 56K Sportser Linux-Mandrake 6.1 or
6.5 PowerPack
Get this sucker working first and then go SCSI or whatever
ya want. -jctwu
\_ ABIT BE6-2. i740 is trash. I'm pretty sure the ABIT is
an ATA/33 MB so the ATA/66 is a waste. Don't get ISA
cards for anything. What you should really do is decide
how much you feel like pissing away on a computer and get
the most bang for the buck in areas where you need it most,
be it video, sound, storage, cpu speed, etc. You may find
you don't even want certain things such as a soundcard or
modem if you're building a networked server box.
\_ the BE6-2 has one ISA slot. Both BE6's support
UDMA/66. The speed gain from not having any ISA cards
sounds reasonable, but is unproven, and I'd choose
2 ISA slots today. As for modem and sound card,
I'm not assuming the reader is so dumb to think they're
required cards. Also, I've had good experience
with the BE6 personally; have you had the same with
the BE6-2? The i740 is $35; I've got 3, and they all
work fine. Please tell me why it's trash. It's hard
to find a well-supported accelerated video card, but
I know RH Linux 6.1 supports it. As for deciding on
bang for the buck, that sounds all well and reasonable
so why am I providing the list above? Because they
all work with Linux, are fully tested, are cheap,
all work with Linux, are tested, are cheap,
and are a good starting point when you could be
all work with Linux, are tested, cheap, and
are a good starting point when you could be
buying high-powered stuff (a new video card) that
doesn't have driver support. -jctwu
\_ go to http://www.eis.com and get a preconfigured solaris x86 box
instead :-)
\_ Uh huh... "Quick! Spend big bucks on the slowest *nix
available!" I'd rather screw yermom.
\_ momma always says, "sometimes, slower
is better". Her customers agree!
\_ Slow on cheap machines. Faster & more stable
than linux on more expensive stuff (SCSI, MP, etc.)
\_ It's Solaris x86. I could run anything really
fast on "faster & more stable" hardware. sheesh.
\_ consider the MICROSTAR MS6120 440BX motherboard, does dual
PII/PIII.
\_ heck, consider the ABIT BP6. Does dual Celerons w/o
weird mods. A friend of mine set up NT with it last
week and it works fine on NT -jctwu
\_ try <DEAD>www.memoryshippers.com<DEAD> . local, great prices on celeron
+ motherboard.
\_ http://www.centralcomputer.com
\_ computer shows have great bargains.. http://www.robertaustin.com
they have Oakland shows every month or so...
\_ http://www.hitech-usa.com or walk over to shattuck ave.
\_ those assholes are the biggest jackasses in the world. pay
big bucks to have people treat you like shit. just go to the
oakland computer show. anything is better than that garbage.
\_ whine whine, bitch bitch.
\_ Do not go to hitech-usa.
\_ Dell also sells machines for more than they're worth
\_ Buy parts and assemble.
\_ Find someone we don't know about.
\_ Figure out what hardware you want, then spec it out and have
someone like Central Computer build it. It generally doesn't
cost any more than buying the parts and building it yourself,
and they'll burn it in for you and make sure it boots. Most
of the small local places that will configure you machines
will not charge you for Windows if you don't want it.
\_ Maybe you should consider FREEBSD or MACOS or another
ALLCAPS operating system. |
| 1999/10/17-19 [Academia/Berkeley/Classes, Computer/HW/CPU] UID:16723 Activity:moderate |
10/16 RedHat's compatibility pages list some sort of problem with
AMD K6-2 300mhz. Has anyone gotten a replacement chip like
they suggest? How much hassle/money required?
\_ Never buy AMD and don't believe the BS that they are faster than
Intel... Read you 61C textbook and the anecdote about the
philosophy that faster was always better,
even if it wasn't always right.
\_ It looks as if the Patterson and Hennessy book is, overall,
very critical of Intel and says nothing about AMD vs. Intels.
\_ I was referring to a quote, I think by Von Neumann, and how
Cray's floating point was fast, but partially incorrect.
\_ I don't think Patterson was too fond of Von Neumann
either. According to Patterson, Von Neumann ripped
off the idea of a stored program computer from two
other researchers at Yale and claimed the term "Von
Neumann machine" for himself. Von Neumann was a
mathematician, not an engineer.
\_ CDC was founded in 1957, and the CDC 6600 was around
1964. Von Neumann died in 1957.
\_ Uh huh.. Intel has a rep for doing it right? Anyway, we're not
talking advanced science apps to predict the future of mankind
(will that asteroid hit?). It's a cheap-ass RH box. As long as
it performs basic function. Don't read a 61C book. Read the
release notes. Not everything in life is an academic issue that
needs to be perfect, follow the IEEE, ISO9000 and 61C standards
of correctness. As to the original question, the hassle involed
is buying a new CPU and possibly MB and installing same. Prices
vary by the minute. Hassle is static.
\_ They don't replace it for (almost) free? That's what
the Redhat page implies
\_ That was a long time ago. I doubt AMD is still
honoring that offer. |
| 1999/10/15-17 [Computer/HW/CPU, Computer/SW/Languages/Misc] UID:16718 Activity:low |
10/15 Can anyone recommend a *functional* freely available word processor
for *nix. WP8 has just pissed the last remnants of a fuck out of
me, and as much as I'd like to use cat | latex | dvips for everything,
I'm not fast enough with latex to make it time-efficient.
\_ It's OK; you're not enough of a writer to make it worthwhile.
\_ StarOffice, from Sun, is not bad.
\_ I'm a big fan. It's just like MS Office (maybe not
such a good thing) but it's free. Seems to run as fast, too.
\_ Get faster with latex. Latex + emacs, the choice of professionals
\_ Latex + text editor IS NOT A WORD PROCESSOR. It is
a way to professionally format documents. That does not
make it a "word processor".
\_ professional whats? most profesional writers(not technical) i
have known use macintosh.
\_ ED! ED! ED IS THE STANDARD!
\_ so who is this Ed guy?
\-if you are a "computer person" and spend most of your time
writing text and not including tables and pictures and such,
emacs+auctex+tex is a great choice. if you are doing graphical
things a lot or are "too short", get some wysiwig. youc an also
use wordperfect-mode for emacs :-) --"psb" |
| 1999/10/13-14 [Computer/HW/CPU] UID:16700 Activity:high |
10/13 12"active Laptop.266 Cyrix CTX.32Mb ram.2 mb video.
$1200 is this a good deal? has anyone had problems using linux or
freeBSD on one of these machines (i.e.wht's not supported?) - paolo
\_ depends on the brand name. Also, I think you have no warranty
you can get new clones for that much - you must check the video
chipset and cross-ref with the X Window system you'll be using
\_ 1200 for a this is terrible. -jor
\_ For 1200, I can get you 3 of these *and* some good swamp land.
This is total robbery. Do *not* do this. Go look at *new* prices
for the same laptop. |
| 1999/9/26-28 [Computer/HW/CPU, Computer/HW/Drives] UID:16601 Activity:high |
9/25 I have an old P133. I just bought a 10GB hard drive in it but can't get
my bios to recognize it as anything greater than an 8GB drive. How
can I get around this?
\_ What the hell crackhead stupidass ED user deleted my reply?
Fine. this whole stupid thread deserves to die. Botton line:
Install linux, or any real OS. problem solved, you now have access
to the full 10 gig.
\_ [restored.. because I can. your linux rant wasn't there]
\_ what the hell for? none of it was useful.
\_ Check if the MB manufacturer has a bios update online. If not,
you'll need a new MB. Nothing else to be done.
\_ Do classic pentium motherboards even have flash bios?
\_ The later ones did. I own one. Earlier ones, no.
\_ Try add on EIDE(ATA33/66) cards, they might work, although
some cards cost about the same as a cheap mobo.
\_ A new high quality BX mb can be had for $99. An old one
for his toy you can get for free from someone else that's
upgrading. The MB is going to have AT-33/66 support.
The dude needs to either forget about the 2gb, or shell
out for a MB made in the last 7 years. This whole thread
is just silly. We're talking about spending money or using
\_ oh yeah, he probably has a socket 5 MB, and P133 is a 90s
product too.
hokey software so he can keep using his socket-3 MB. Let
it go. Join the 90s.
\_ maybe he cant get a used/old one from someone, that's
why he's asking. why dont you donate him a BX board?
BTW, while you are at it, why dont you donate
CPU and memory to go with it? used is fine.
\_ Bullshit. There are numerous places on the net
still selling socket-7 boards. Anyway, why should
I donate anything to someone so stupid that they
didn't perform the most trivial and basic of
research before spending whatever on a device that
their MB doesn't fully support? It's his fucking
fault for being stupid. My email address is not
root@donatestostupidpeople.org.
\_ P133 uses socket FIVE and is a 90's product.
\_ Ooooooh... socket... FIVE!!! Woo woo!!! *laugh*
\_ nothing to brag about, but at least
get the facts stright before ranting away, idiot.
\_ Can this be accomplished with drive overlay software? Or does
overlay software just trick the OS and not the hardware? -dans
\_ Dude, forget it. Stop being cheap. Overlay software is dead
and only works with certain versions of certain operating
systems. The 2gb is nothing. Either get a real MB or forget
the 2gb. You're being silly.
\_ we have the right to be cheap! not everyone makes
100k/year and drives BMW.
\_ you mean people get by on just $100k/yr?
\_ I don't drive German cars. If he can afford to buy a
10 gb drive without even researching to find out if his
system will fully support it, he can easily afford another
$50 on a MB made in the 90s. |
| 1999/9/17-18 [Computer/HW/CPU] UID:16541 Activity:high 66%like:15848 53%like:17542 61%like:17733 |
9/16 Software jobs available at Intel. See /csua/pub/jobs/intel
-emarkp
\_ How much is your ref. bonus?
\_ Emailed received from Intel upper middle manager signed off with
the phrase, "Have a productive day, Joe". No thanks.
\_ What does this even mean? If you're trying to make a
statement, it would help if you constructed a valid
sentence. I'm in the group that's hiring. I like the
group, and put the job listing up as a favor. We're
looking to extend several job offers before the end of the
year. If you send a decent resume, you'll probably get an
interview (though I recommend using correct or at least
reasonably decipherable English on your resume). -emarkp
\_ A bit touchy? Check with your HS english teacher. That
\_ only if the e-mail signed off, not the manager
try "manager who signed off..."
is a valid sentence.
\_ only if the e-mail signed off, not the manager try
you're totally on target about the grammar, that's
entirely unrelated to the point, namely that Intel
is a slave ship. ROW!... ROW!... ROW! When you hit
your early 30s and they fire you to make way for 5
NCG's, don't come crying to the motd about it.
"manager who signed off..."
\_ That's incorrect. "manager" is the object of the
VP "signed off". Don't make me get my grammar
books out and go balistic on your slave ass.
-Straight A English Major
\_ "Emailed received"? I don't think do.
\_ Sure is. You're simply wrong. Sorry. Even if
you're totally on target about the grammar,
that's entirely unrelated to the point, namely
that Intel is a slave ship. ROW!... ROW!...
ROW! When you hit your early 30s and they fire
you to make way for 5 NCG's, don't come crying
to the motd about it.
\_ That's a lovely myth. Come back when you can
think for yourself. -emarkp
\_ http://www.faceintel.com Some time soon you might want
to start thinking for yourself instead of letting
your manager do it for you. Or maybe these
people are *all* just liars and nincompoops.
Yeah, that's it.
\_ Actually, the stuff in faceintel reads
like tabloid. I've never seen or heard
of anything like this in my department.
Obviously I can only speak for my
department, but it's remarkable this
nut feels like he can speak for the
entire company. -emarkp
... The second sentence is accepted slang.
I have a decent resume but don't need an Intel slave job. I
suppose having Intel on your resume as a NCG is good but I'm
too old for that crap. I don't need it. Thanks.
start thinking for yourself instead of letting
your manager do it for you. Or maybe these people are
*all* just liars and nincompoops. Yeah, that's it.
\_ What is it with the "corporate slave" stuff? I get
paid well, I develop software I'm proud of, I go
home at 5pm, and I get regular raises, bonuses, and
stock options. Furthermore, the belief that Intel
will fire you after a few years for a recent college
grad is a myth, at least in my department. -emarkp
\_ http://www.faceintel.com Some time soon you might want
to start thinking for yourself instead of letting
your manager do it for you. Or maybe these
people are *all* just liars and nincompoops.
Yeah, that's it.
\_ I don't trust Intel, but I don't trust a site
either. |
| 1999/9/1-2 [Computer/HW/CPU] UID:16445 Activity:high |
9/1 Wow, a 500 MHz Apple computer can run PhotoShop operations 2x
faster than a 600 MHz Pentium 3. What is a Velocity Engine? That
programmable CPU instruction thing or some superscalar thing?
\_ If you check Adobe's web site, there's a patch for Photoshop 5.0
\_ 32?
that has optimizations for P3 processors. Now recheck those
benchmarks again.
\_ oh duh, it's just Apple consortium's answer to the P3 instructions;
I wonder if there's a KNI plug-in for PhotoShop ...
\_ This isn't an Apple chip. This is a Motorola chip. Apple is
not a chip design/fab company like intel/amd/hp, etc.
\_ yes yes, the Apple 'consortium', offhand; goddamnit, I hope
everybody isn't taking the behavior in the vaporware
discussion below to be the norm. I sure hope the hell not.
Jobs, just like the last time when the G3 was twice as twice, etc
\_ But it is the norm. tom is involved.
\_128 bit processing vs. 64 bit for current pc chips
\_ 32? isn't Merced 64?
\_ merced isn't "current", it's vapor
\_ merced is silicon.
\_ It's not available to the
public. You can't buy one.
It isn't typical to call something that _/
is an actual manufactured part "vapor." It is shipping
engineering samples. This is standard in the IC industry.
Typically it means 6 mo. to 1 yr. before consumers get the
part.
\_ On the contrary, it's customary to call anything not yet
publically released as "vapor" - so much can change until
it's released, nothing before really matters.
\_ you're making up your own definitions. not even
idiots believe that. it exists. not even an
idiot believes it won't ship.
\_ it's only customary among idiots. -tom
\_Is this benchmark real this time instead of an outright
fabrication like with the G3?
\_ benchmark exploitation isn't a new thing. take matrix300
for instance. people were boasting 60x improvements.
you can read all about it in the p&h book.
\_ This is a "Steve Jobs Demo Benchmark". I'll believe it when
it's confirmed by multiple sources outside Steve's Reality
Distortion Field. In the mean time, it's just more bullshit from
Jobs, just like the last time when the G3 was twice as fast, etc
which it wasn't. Steve Jobs has a Microsoft quality record
regarding truth in advertising. Not impressed. Even if the
speeds claims are real and apply across the board instead of a
single specific benchmark, the OS is still shit under the hood.
Stick a *nix on it and it might be ok.
\_ MacOS X is coming. linuxppc is here. -tom
\_ "is coming". vapor.
\_ Well, no, actually, MacOS X Server is here, I'm
running it. -tom
\_ Make up your mind. Is it coming or is it here?
Can I walk into compusa and buy it? Can I order
it from Apple's website?
\_ Probably, and yes. The client version isn't
here yet, but for someone with such obvious
brilliance, the client (which includes the
new Carbon API) isn't necessary. Sign your name,
idiot. -tom
\_ Sorry, being an idiot, I forgot to sign. You
know me as... --your name, idiot. -tom |
| 1999/8/28-30 [Computer/HW/CPU] UID:16409 Activity:kinda low |
8/27 What do companies like intel or mips do with all those defective
chips or wafers that never make it to the shelf. I want one
just to hang up on my wall. I'm sure Intel gets low yield with
their one acre sized chips.
\_ Intel made nice keychains out of the mistakes when they first
starting fab'ing pentiums. I think they sell the defective ones
as geek jewelry now.
\_ You can buy them at the Intel store in Santa Clara:
http://www.intel.com/pressroom/archive/releases/Cn72399a.htm
or online at <DEAD>www.adviceforpcs.com/uniquelyintel<DEAD>
\_ I have a full wafer on my cube wall, but I can't take it out
of the building.
\_ Dilbert boy. |
| 1999/8/20 [Computer/HW/CPU] UID:16349 Activity:nil |
8/19 There are a lot of spys/espionage in companies like AMD, Intel,
and other hardware companies. But what about software companies?
Why aren't there more crime associated with stealing software
intellectual property?
\_ why do you think there is less crime? where are your #'s for
HW vs SW espionage? |
| 1999/8/18-20 [Computer/HW/CPU] UID:16338 Activity:low |
8/19 Can somebody clue me into dirt-cheap refurbished note book
dealers on the net? I am looking for something that
I can haul around with me to the library. I do serious
computing at home on the desktop. I need a glorified
word processor/telnet box for travel and the like. -fab
\_ what's your definition of dirt cheap? I got a micron
refurb P5-266/64MB/3.2G/13.3TFT for $1200 from PC Nation.
Adequate for most daily browsing/word prc/telnet. 13.3TFT
was very important to me, and this was the cheapest I found.
Their stock was limited, I got mine about 3 months ago.
\_ Since I desire only to use Word and Telnet, I'd be willing
to stoop to some machines with 486 processors just to save
some money. Dirt cheap is < $1000. -fab
\_ Word is about the biggest RAM/processor hog you'll find. -tom |
| 1999/7/5-6 [Computer/HW/CPU] UID:16071 Activity:very high |
7/4 Does anyone know how processor-intensive MP3 decoding generally is?
My Celeron 300 is claiming to consume only 0.1% of the CPU according
to top but my PentiumMMX 150 could barely keep up.
\_ this doesn't answer the question, but I bet FPU, MMX, OS, and
player all could have contributed to your results
\_ pMMX150 256k L2 win95 winamp 3 >50% CPU vs.
celeron300A 128K L2 linux 2.2.5 xmms 0.9 ~0.1% CPU
I don't think this should account for a 500x speedup.
\_ I do. A guy I know swapped out a 200MhZ PPro for
a PII overdrive, his FFT benchmark got over a 200%
increase in speed. --jon
\_ Celeron 450 (300) Win98 10-15% CPU
\_ WTF?!? I get like 1%-3% on NT for OC 450
\_ Amount of memory and types/sizes of caches can also be
fairly vital.
\_ Could be that the tools you're using to measure are wrong. |
| 1999/7/2-5 [Computer/HW/CPU] UID:16062 Activity:high |
7/02 Does anyone know how to find out the clock speed of a processor?
For example, you can find out how much memory a computer has by
using top, but how do you find out how fast a processor a computers has?
\_ reboot and check the bios.
\_ On SunOS 4.1 there's a "sysinfo" utility. Don't know about other
\_ dinfo
versions of UNIX. -- yuen
\_ sysinfo is a freeware program, not part of SunOS
http://www.magnicomp.com/sysinfo
\_ http://home.earthlink.net/~bhami/rosetta.html
includes: lscfg, uerf, dmesg, ioscan, hinv, prtconf
\_ On Irix, "hinv"
\_ Many PC Unices (FreeBSD, Linux), that stuff will be somewhere in
\_ The only thing in dmesg which looks relevant is the following:
/proc, but just running "dmesg" and reading it is usually easier
does that suggest I have an 80 Mhz Machine or something
completely different?
\_ bogomips are different functions f(clock) for each cpu type,
e.g. pentium, pentium II, 486, alpha, sparc, are all computed
differently. check /proc/cpuinfo on linux for real measured
MHz.
--dbushong
\_ The only thing I found in dmesg which seemed appropriate was:
Calibrating delay loop.. ok - 79.67 BogoMIPS
does this mean my machine runs at 80 Mhz?
cat /proc/cpuinfo --dbushong
\_ Sounds like you're on a Linux box (mention this next time,
will save everyone a bunch of time).
"cat /proc/cpuinfo" will tell you the clock speed, and no,
bogomips do not correspond in any way to MHz --dbushong
\_ psrinfo -v on Solaris |
| 1999/6/16-17 [Computer/HW/CPU] UID:15969 Activity:high |
6/16 http://www.cnn.com/TECH/computing/9906/15/supercomp.idg They claim 60,000x faster than a PII350 for $1000, next year \_ This probably is bullshit. This claims to be a reprogrammable FPGA which gives you the advantage that a special purpose processor would (like an mpeg decoder chip or a voodoo chip) but even those special purpose chips will give you a performance boost over a Pentium by only a factor of about 20x not 60,000x. \_ How is their technology different from the TransMeta stuff? Sounds similar... \_ "It will be some time before people will believe this is real. But in six to eight months, it will be an even bigger story, when the applications roll out" - I'd like to see an application roll out in six to eight months! \_ Self-modifying microcode on a Dandelion. Been there, done that. \_ Since Intel first invented the microprocessor, every year there has been a new bit player that comes on claiming to beat Intel. And every year they have failed. This one will be no different. \_ you are an idiot. people regularly make this claim, and reguarly, their processors are faster. they just don't sell as much. unless you mean "beat intel" as in "make more money than intel." -ali. \_ "beat intel" as in having the most bang for the buck: performance / price. The best technology that people can't afford is meaningless. So what if DEC alpha is the fastest? How many people can enjoy it? Intel has 6-7 billion-dollar- fabs around the world and they consistently get the best yields. That's why they're the world's #1 supplier of chips. \_ You have no idea what anyone's yields are. You're making shit up. Yields are not public knowledge. You're full of shit. \_ wooptie fucking doo. memory bandwith will be the probelm for multimedia, not bit flipping power. besides, this processors is not meant as a general purpose desktop cpu. -ali \_ Don't care what its meant for. Can it be used as one? |
| 1999/6/15-18 [Computer/HW/CPU] UID:15961 Activity:nil |
6/15 Want to buy (cheap) a standard Pentium motherboard which I can
plug a P75 and a few old SIMMs into. Please mail randal. |
| 1999/5/25-27 [Computer/HW/CPU, Computer/SW/OS/Windows] UID:15874 Activity:high |
5/25 Does anyone know any good (and preferrably free) x-server for
Win 95/NT?
\_ <DEAD>www.uk.research.att.com/vnc<DEAD>
it's free and the whole thing can fit on a floppy. I haven't
used it, but have heard good things about it.
\_ an X server VNC is not. But its still pretty cool. -ERic
\_ Hilarious, slow, crashes really creatively (even for me) -John
\_ Use 8 bit graphics and stop trying to play netrek over it.
If the server is a unix box, it works great, the NT server
puts a huge cpu load on the host due to lack of open NT
source leading to much guessing on the developer's part.
Crashing? RTFM, John. NT has multiple screen modes.
\_ I dont know how good it is but here's a free one:
http://aquarius.franken.de/software/net/X/MI_X/mix_faq.html -ERic
\_ Exceed from Hummingbird, but it costs $
\_ it's free on campus/to students
\_ anon ftp http://tuna2.berkeley.edu/pub/depot/x513wnt.exe
\_ depot has 6.x versions online which are pretty
much the same as 5.x but cleaned up a bit. More
of a 5.2 than a 6.1 but I'm not working in the
Hummingbird marketting department.
\_ X-Win32, available from http://www.starnet.com -- yuen |
| 1999/5/24-25 [Computer/HW/CPU] UID:15869 Activity:kinda low |
5/24 Roadmap to what your next computer should be:
http://setiathome.ssl.berkeley.edu/stats/platforms.html
\_ Hardly. They're rating a p5-60 the same as a p3-550. This is
worthless as a power measurement. It's all averaged in. Go take
Stat 2 or Math P or something like that.
\_ That "rating" also doesn't take into account (in the PC case) how
many processors your PC has. On my dual Pentium NT machine it
uses both CPUs (even though the FAQ says it can't) and reports
10 hours CPU time per work unit, whereas if I limit it to run on
one CPU it reports 20 hours CPU time per work unit. -- yuen
one CPU it reports 20 hours CPU time per work unit. By the way,
David Anderson himself uses an iMac. -- yuen
\_ Many OS'es report any intel-arch chip as "i386" so you can't
tell from that chart which are 386/486/Pentium/PPro/K6/K6-2/etc.
\_ Yes, but how fast does it run Microsoft Word? |
| 1999/4/1-2 [Computer/HW/CPU] UID:15679 Activity:high |
3/31 I'm running irix on a two processor machine. Is there a way
to lock a certain process down on one processor
(in the sense that process A can only run on processor 1 and
not on processor 0) ? Lets assume I have root.
\_ I'm just surprised that a berkeley student managed to get his
hands on a dual processor SGI. Those things are freaking expensive.
\_ Actually, I dont. I'm running simos (processor simulator)
in a dual processor mode while running irix5.3 on it.
\_ Dual processor Octanes only go for about $30,000 . It's
not inconceivable that he has a job or is working on
a project that bought one. Of course, we now know this is
not the case. --dim
\_ For $30,000 you could have gotten several of their new
540 and 320 series workstations. From what an SGI
engineer told me those machines have a considerable
amount more bus bandwidth but cost much less than their
O2's. And the best part is that the latest linux kernels
support them.
\_ or spend $10,000 on a dual (or quad?) pentium,
and run solaris on it.
\_ Why would anyone spend $10,000 on crap from Intel.
\_ Because I'm not a religious fanatic?
I just picked up a dual 604e 200 card ($149) and
the combined total of 802 BogoMIPS puts even dual
333 Pentium II (~$1000) with 599 BogoMIPS to shame.
x86 is a dead arch., and with the latest offerings
it is starting to stink (like putrid meat).
If you need SMP for business reasons go with
UltraSPARC II.
\_ uh, "bogomips" says absolutely nothing about how
fast the machine is. -tom
\_ if it's pure integer, ultrasparc is a total waste,
price-wise. Now tell me the specrate on your
dual 604e system. Then give price of *system*.
\_ the NT machines have "0 cost" texture mapping but
don't claim to beat the polygon rates of our octane or
origin+IR for realtime VR (cave). my $5000 dual
450 MHz PII, ultra-wide SCSI, 256MB desktop w/ linux
runs large autoconf+cc loads comparably to the octane,
and snappier X console than any of our ultrasparcs.
\_ % runon n cmd
runs cmd on processor n
% man sysmp
tells you how to lock a process down to one processor,
and how to restrict a processor to only certain processes. |
| 1999/3/20-22 [Computer/HW/Memory, Computer/HW/Display, Computer/HW/CPU] UID:15617 Activity:insanely high |
3/19 Dell anounces it will ship PCs with RedHat Linux preinstalled:
http://www.news.com/News/Item/0,4,34036,00.html
\_ Is there going to be a Linux refund day?
\_ Why? Isn't Linux free, unlike Windoze?
\_ Most of it is free but RedHat throws in a bunch of
commercial crap that nobody uses (like a Real media server)
so that can charge $40 for it. It also comes w/ a manual.
\_ RedHat's Lunitux. And it costs them money to install it.
\_ Good luck to Dell to get Linux users to buy Dell computers.
\_ What's wrong with Dell's? They only cost twice as much as a
computer you build yourself.
\_ and they're worth more than twice as much. -tom
\_ The only x86 based PC worth paying extra money for
performance are IBM's and SGI's.
\_ Yeah all that off the shelf hardware magically becomes
much more reliable and better performing when installed
by a Dell technical genius.
\_ No, it becomes WARRANTIED, and YOU DON'T HAVE TO SPEND
TIME PUTTING IT TOGETHER AND MAKING IT WORK. -tom
\_ My parts are covered. I think I can afford an hour or
two to save $1000 or more. Anyway, it's a hobby thing
and I get *exactly* the hardware I want. To the person
below, who said I was using cheap parts? What's a cheap
part anyway? They all come from the same factories.
Please name a "cheap part" that will be hard to install.
You just need a good case. The rest is generic.
\_ Selecting a mother board would be your first
problem. Intel, Asus, SuperMicro, Tyan, etc?
For example, I tend not to want to use a
model that just came out. New models tend
to have problems fixed only in later revisions.
They are usually only tested with brand name
components, and even that is not guanranteed.
When there is a problem, you are sometimes
left wondering whether the problem is with
the PCI bus, a BIOS setting, a motherboard
switch setting, your network card, a device
driver problem, compatibility problem, etc.
When going crazy trying to solve a problem,
you may unknowingly create a second problem
zapping another component with static.
I would gladly let Dell do the dog work and
quality control for me, and let them iron out
all the problems.
\_ I don't see MB selection as a "problem". People,
this isn't rocket science. You're just assembling
stock parts. This isn't the old bad days of Dos
2.1 (or even 6.2) where nothing worked and it took
hours of fucking with config.sys to get something
working. The parts Dell uses are the same you're
getting. They don't come from a Magic Dell
Factory whever everything is always wonderful
Factory where everything is always wonderful
while your parts are coming from Factory Hell
where nothing ever works. It's the same stuff.
The only difference is you won't be paying for
parts you don't need/want and can put that money
towards things you do.
\_ Yes, Dell parts are the same you are
getting but they do a couple of things.
One, they test and throw out all the
defective parts. Two, they make sure
all their components work together
well. Three, they configure and
assemble the system for you on a
production line. Four, they test the
whole system again after assembly.
Go to a computer reseller, open a
box of say Diamond Multimedia video
cards or Intel motherboards and test
them. The defective rate can be as
high as 10 percent.
\_ Ok, granting all these things are true,
pleasure.
which they're not entirely, even if I got
a defective board, so what? I return it.
I get a new one. This isn't for work. It's
a hobbyist home system. Dell buys in bulk
so they have lower price purchase prices
but their rates aren't *that* much lower,
you *are* paying for them to build it and
run their business and you *don't* get
exactly the parts you want. If you don't
care about what's really under the hood,
then you are a Dell customer and I hope
the value you feel they're adding is worth
the price you're paying. Nothing wrong with
that. It just isn't the best you can do for
your money.
\_ Just make sure you test your
motherboard well before assembling
your system. I hate taking a
motherboard out. I also hate
driving one hour to Hi-tech USA
(or whatever) to exchange a part,
or packing stuff and going to the
mail shop.
I do care about what is under the
hood, and Dell components seem
fine to me.
Price advantage of resellers over
brand name systems has been eroding
over the past few years, so I doubt
you can save much money.
You do get to use the exact parts
you want though, and if you need
to tinker with your system in the
future, you already know it
through and through. Enjoy building
your system. I love the ones I
built, but would not do it again.
\_ look, have fun building your box, but you're fooling
yourself if you think it makes economic sense. -tom
\_ Whatever Tom.
\_ Ja, building your own system is a nice learning
experience, but after doing it for a few times,
you decide that the money saved is not worth the
amount of effort you have to put in. You also
realize that the cheaper the components you used,
the higher the amount of time you have to put in.
\_ Agreed. I've been building my own computers since the
486 days. But for my dad, I'm gonna plop down some $$$
for a Dell Celeron. It's just not worth the hassle
each time.
\_ celeron sux. go with amd-k2.
\_ On Friday I saw a celeron 300MHz on an
overclocked board (66->100Mhz) perform a
little less than twice as fast as an AMD-K2
300Mhz. Celeron used to suck, but no longer
and for $60/chip... do the math fuckwit.
\_ I built my first computer before the 8086 days. I
don't get so easily "confused" that I need Dell to
pick all my parts for me and charge me more for the
pleasure. For your dad, sure, spend his money.
For yourself... don't you care?
\_ how much do you care about your time? I can
configure, price and order a Dell system in 5
minutes, sitting at my computer. How much time
do you spend contacting 8 different vendors
for parts? How much time do you spend putting
it together? If one of the DIMM slots is bad,
how much time do you spend taking the machine
apart again, and trying to convince the motherboard
manufacturer that it wasn't your DIMMs that were
bad?
If you're still a student, or some other class of
person whose life decisions are more cash-bound
than time-bound, that's fine. If you have a job,
and particularly if you're a computer professional
in a job that pays good money and probably takes
more than 40 hours, why the hell would you spend
one minute more of your time than is necessary,
to try to wring, at best, minor cash savings out of
a purchase? Seriously, how much does your free
time cost, per hour? -tom
\-that's a psb law: "there are two kinds
of people. peopel with more $ than time and
people with more time than $."
\_ It's a hobby, you freak! How much does your
netrek or other gaming time cost per hour?
\_ I don't try to claim that netrek is
cost-effective. Like I said, if you enjoy
building systems, have fun. -tom
\_ Well, nowhere in my answer did I say I was gonna
stop buying for myself. Unless a computer maker
magically has all the exact parts I want and
I don't feel like spending an entire weekend
going to all these small stores finding the
cheapest prices and constructing it myself.
\_ It isn't about money. I'll spend a few extra
bucks _AND_ the time to get _exactly_ the system
I want. _Exactly_ what I want. Not from their
approved list on a web form. I'm glad the
consumer route works for you.
\_ Don't some companies, even Dell, allow
you to customize the computer you buy
from them?
\_ On some web form, yes, to a *very* limited
extent. But if you want that safe feeling
from knowing you have the same computer
Dell has built for 10k others before you,
you can't get that from a highly
customised system. It isn't possible.
At that point, you're now paying them to
collect and assemble the same random
pile of parts you would've built for
yourself anyway. You've already done
research and everything else short of
actually popping it all in case. Might as
well just take that last step yourself.
And btw, I've seen vendor specific
versions of some hardware which require
you to use the vendor drivers. Generic
drivers don't work. Upgrading sucks too
because the cases are fucked also. Caveat
emptor and all that no matter which way
you go. Dell is not the be-all end-all
perfectly safe answer to hardware buying.
\_ Rather the contradiction with "off
the shelf" parts & "vendor drivers",
no? In any case, buying from a vendor
is all about convenience and no hassle,
especially when you need 20-30
identically set-up computers.
\_ No, it isn't a contradiction.
They're only tweaking the HW to
force you to come to them for future
support and upgrades, not to enhance
it in any way. Compaq is especially
prone to this and has a lousy system
of storing and categorising patches
to go along with it. Other than
making sure you're trapped with them
forever, there's nothing better or
different about the HW. It's just
a scam, not an improvement.
As far as 20-30 goes, go ahead. Buy
Dell. No one said you should build
20-30 machines from scrap. The idea
is stupid and you're creating a weak
straw man argument. |
| 1999/3/17-18 [Computer/HW/CPU] UID:15603 Activity:high |
3/16 What ever happened to Cyrix?
\_ who?
\_ Bought out by someone (Nat. Semi?) but still making lame chips
on IBM's fab. Both AMD and Intel are faster, more reliable and
dirt cheap (for Celeron class), so what good is a cyrix?
\_ Yes bought by National Semiconductor. Their chip is
powering (if you can call it that) the $299 computers.
Most agree that Cyrix's Pentium-class offerings were just
fast 486-style processors. Not so sure about their latest
offerings.
\_ Didn't Cyrix announce plans for the "computer-on-a-chip"...?
kind of like on-motherboard video/sound chips but even more
evil & twisted. -jctwu
\_ WinChip or something like that. integrated cpu + video/sound.
\_ They already made it a while ago.. called MediaGX --dbushong
\_ That sounds right. It was pretty much a bomb, yes?
\_ and then there's SiS's PCI+IDE+graphics+sound on a chip.
\_ Winchip is IDT I believe. |
| 1999/3/3 [Computer/HW/CPU] UID:15529 Activity:high |
3/3 What are the cons to overclocking say a Celeron 300 to 550 Mhz?
\_ Celerons are clock-locked at 4.5.
\_ Celeron 300As are clock-locked at 4.5.
450 ~= 4.5 x 66 bus
550 ~= 4.5 x 125 bus
PCI bus is 33 MHz, but the board only comes with 1/2 and 1/3
dividers. 125/3 = 42 MHz, ~ 25% over-stress on PCI cards. Most
likely this will fail because one or two PCI cards will choke;
graphics cards heat up, too. Christ, 450 is fast enough.
\_ What's a good 112 MHZ mainboard? If the board manufacturer
knows that the divider won't line up with PCI standards
why do they release it? Why can't dividers be differant than
1/2 or 1/3?
\_ people think "oh, it can support 112/125/133" it must
be cooler than this one that's "slower". Most PCI cards
are ok up to 112 MHz (1/3 divider), ~10% stress. I heard
the new Coppermine is going to do 133 MHz bus (1/4 PCI
divider) -- the divider things are hard-wired and are
expensive to make variable or to put more on
\_ dual celeron 400 over-clocked on 75mHZ board gives 450mHZ
how comparable is this to a pII450
\_ is this a troll? |
| 1999/2/26-3/30 [Computer/HW/CPU] UID:15489 Activity:moderate |
2/26 AMD outsells Intel in Janurary. Kewl.
http://www.news.com/News/Item/0,4,32972,00.html
\_ Of course, if Intel sold its products below cost (like AMD
is), that might not be the case.
\_ its all about market share man. Do you think MSIE would be
where it is now if microsoft wasn't giving it away Free?
Hell in some cases they're paying people to use it!
\_ Which is strange--MS is being hammered, and AMD is
being praised, even though they lost $100M last year
and are losing money with every chip they sell. If
Intel did the same thing, they'd be attacked just as
MS is. And now, AMD comes up with the K6-III, which
will lose them even *more* money per chip (and they
don't really want to sell many of them). Whatever.
\_ They only lose money because they have terrible
manufacturing processes. Intel doesn't make any
money on the low end chips either. MS gets hammered
because they have a monopoly. AMD is forcing intel
which basically had a monopoly ot lower its prices.
\_ Actually, you have to design for manufacture as
well as manufacture well. Yes, Intel does make
money on the low end chips (trust me). It's
just not as much as on the high end chips
(obviously). AMD has stated that they have to
achieve an average selling price (ASP) of about
$100 to be profitable. Currently they're
selling at *much* less than that. It's the only
reason they're selling anything at all.
\_ How long can they stay afloat by losing money
this way? Intel after all has much more cash.
\_ AMD has other not-as-high-profile
products that we never hear about
because the news media doesn't think
that those products are sexy enough.
Perhaps those are the products that
are keeping AMD afloat. Besides,
Amazon isn't making much money and
their stock is (or was? I don't
remember) worth rather a lot. Success
is defined by something like the
Calculus of FITR, and I'm glad nobody
relies on me to make a critical
business decision that will increase
business success. Business is all a
load of goat droppings anyway, and we
should all go live on farming cooperatives. |
| 1999/2/26-3/30 [Computer/HW/CPU] UID:15486 Activity:nil |
2/25 What's the difference between the Celeron 300a PPGA and slot1?
How good is ppga for overclocking?
\_ ppga is supposed to be $10 cheaper for Intel to make, but it
ends up costing the same to us. Last I heard they discontinued
Slot 1 300As. ppga goes into a motherboard with a ppga socket, and
these motherboards aren't as common and often have a sucky
chipset and are otherwise below average. As for overclocking, I
heard all 300As suck at overclocking nowadays (ymmv). Don't buy
ppga, don't overclock either type of the current batch of 300As -
Get a Celeron 400 Slot 1: I put together two computers for a
business with this and they're running dandy. My boss at my new
job has been nice enough to let me put together my new workstation
and it's going to have this CPU in it. -jctwu |
| 1999/2/12-14 [Computer/HW/CPU, Computer/Networking] UID:15408 Activity:kinda low |
2/12 I want to run a comercial web server. After being bounced from
ISP to ISP because we are using too much CPU time, we are
thinking about getting our own server. Can you comment on
server colocation vs. DSL? (no money for frame relay now) Thanks.
\_ If you colocate, you don't have to worry about the network
connection--someone else maintains it for you. It probably will
also be a faster connection (usually 100 megabit out to at least
a 45 megabit shared connection, vs 384 kilobit for cheap DSL).
Plus the network will likely have fast response, 24-hour support;
if your DSL goes down at 6PM on Friday, how quick will Pac Bell
deal with it? The tradeoff is that you don't have easy physical
access to the box if something goes wrong with it, and colocating
is probably more expensive. You also get an outgoing fast net
connection with DSL; you have to pay for one if you're
colocating. -tom
\_ colocation costs about 30% more than DSL, but you raised
good points about the service of phone companies we
all learn to live with. Thanks!
\_ must be a porn site.
\_ Of course, since we all know how much CPU a porn site must
burn. As opposed to bandwidth. Sigh.
\_ What, network/hard drive I/O doesn't take CPU?
\_ cgi porn?
\_ MP3 live broadcast of acts. |
| 1999/2/12-13 [Computer/SW/Languages/Java, Computer/HW/CPU] UID:15403 Activity:high |
2/12 Most people in the industry are content at the fact that they are
1)in a prestigious corporation 2)pays well 3) good benefits.
How PATHETIC.
\_ pathetic attempt @@ flame bait (though I guess that means I
fell for it)
\_ beats working for a lowly company that pays terribly with no
benefits.
\_ You're right. We should strike for daily blow jobs.
\_ If they hire outside people, yes. I wouldn't want a blowjob
from anyone that works at my company.
\_ I'll join the bait eaters: Uhm, gee, I think you just posted 3 of
the 4 reasons for working anywhere. You had some other reason one
might work in industry that you think isn't pathetic? Or you'd
only have respect if _everyone_ went on to get PhDs? Then CS would
be a purely academic field and no one would own anything with a
chip in it because chips wouldn't exist outside a lab. And who
pays all these super geniuses to do pure research forever? Sigh...
Feeling bad that your cop-out friends all drive expensive cars and
can afford to eat better than rice every other night while you're
kissing your Prof's fat pot sticker filled ass everyday? Bitter
yet? And how about those of us in industry making those big bucks
who never even took a CS class? hehe...
\_ True, but people who usually go into a high tech job for the
big bucks usually ends up producing worthless junk, copying
other people's ideas, suing other people because they can't
come up with their own ideas, making faulty products, or
going along with the fad whatever it may be that day. People
who don't take academic seriously are usually the uncreative
types. Plus, if it weren't for the PhDs like Dave Patterson,
Ken Thompson, John Hennessy, and so forth we wouldn't have
things like RISC chips, chips that pipeline, superscalar and
vector processors, time shared multithreaded multiprocessing
operating systems, and so on. So if you think the academic
world is so bad why don't you just skip college, go straight
into your 'computer programming' company and tell your
employer how much you hate PhDs and universities that do
research in electrical engineering and computer science. |
| 1999/2/12-13 [Computer/HW/CPU] UID:15401 Activity:high |
2/11 Monday is Windows Refund Day in Foster City, for non-Windows users
who received Windows bundled with a computer and have _never_ used
it. (Not even once.) You must also have your original media,
manuals, and certificates of authenticity. Or come along to watch.
http://linuxmafia.com/refund
http://zork.net/refund -- schoen
\_ tuesday is firestone refund day. For everyone who ripped the
tires off their honda accord before they drove it off the lot,
we're going to go to a rubber tree plantation in cambodia to ask
for a refund.
\_ If your tires came with a Microsoft Windows EULA preventing
you from decompiling, disassembling, or reverse engineering
them, selling them separately from the car they came with, or
switching one of the tires to another car -- and allowing
your license to use the tires with that car to be terminated
by the tire manufacturer in the event of breach by you of this
license (in which case you must destroy the tires), and it
happened to say in this "tire license" that you could return
the tires for a refund if you disagreed, then, sure, why not
return the tires if you've already got better ones? -- schoen
\_ would you go to firestone, who made the product, or to the
car dealership where you paid for it?
\_ A good question. People have been asked to call their
OEMs in advance. With only two exceptions in the US,
the OEMs have disavowed all knowledge of the EULA. :-)
Thus, we will ask Microsoft what's up. -- schoen
\_ So what do you want to bet that the microsoft offices will be
observing Presidents day this year, and be closed? -ERic
\_ Good thinking, but they won't: http://linuxmafia.com/refund/faq
\_ How likely is it that you've never used it at least once if the PC
was initially installed to automatically boot into Windoze? Unless
you already knew a long time ago that there's going to be a refund
day.
\_ Because being a die hard linux geek, you inserted your linux
boot floppy the first time your powered it. After all what
use would a true linux geek have for window, pre installed or
not? If you booted windows, you're a loser and don't deserve
a refund. In fact, it wouldn't be a legal request.
\_ Also, apparently some of the newer bundled systems ask you to
click "I Accept" before they allow you to finish booting
Windows. So you can say "I Do Not Accept" and it reboots or
some such thing. We'll see how many people bothered to do
this... -- schoen
\_ Apparently, some companies are forgoing Windows all
together. People who buy low-end SGI's and high-end HP's
were complaining about being forced to use Windows so
now both manufacturers are offering RedHat prebundled.
\_ You're funny! - Intel employee
It makes sense with the multiproc SGI's since Windows
can't scale worth a damn.
\_ I heard a rumor that Compaq let you choose? Sounds
fishy. -John
\_ Because of their close ties with the Intel Merced
development, HP is also apparently working hard to
get Merced support for Linux which should be pretty
interesting. Anybody know the ship date of Merced?
\_ Merced is a piece of shit. By Intel's own accounts
it will "maybe" be faster than the 32 bit offerings
they have during it's lifetime. They're just releasing
it so people will start writing IA64 code until
McKinley (the next next chip) comes out.
\_ Merced does not suck, regardless of the speed of the
first implementation. RISC is reaching its limits,
just like CISC before it. -muchandr
\_ Uh, if CISC is dead and RISC is too where do you
prupose we go now?
\_ multithreading! -nick
\_ Things always take longer than you expect them to, even
if you take into account Hofstadter's Law
-- Hofstadter's Law
\_ Other people are working on Merced, too. Linus
talked about Intel <--> kernel developer contacts
at the Taos "Future of Linux" panel, and said he
expected Merced to be supported upon its release.
\_ Is the Merced backwards compatible with the i32
chips? It would be nice if Intel did what
Motorolla did with the PowerPC chip and ditch
the whole backwards compatibility idea. Intel
chips are too bloated and complicated. Any
architect who thinks that variable length CISC
instructions will last forever needs to get a new
job.
\_ PowerPC was backwards compatible with
the IBM POWER RISC instruction set.
Motorola didn't design PowerPC alone.
\_ Yes, but that unit is almost a separate CPU on
the same die. I think Intel will be able to throw
it away once there is enough Merced soft around,
-muchandr
\_ No one is going to buy Merced, esp. because
it won't be available at resonable prices.
In any case UltraSparc III will be well
established by then, offering greater
performance. And TRUE Linux geeks know that
even UltraSparc pales in comparision to
the ultimate processor that Linus is working
on at TransMeta.
\_ didn't pay much attention in the 90s, eh?
Alpha had greatest performance but everybody
bought intel anyway. there are many factors
besides technical merit or value in the
microprocessor market. SGI is expected to
use Merced in the followup to the Origin
series of supercomputers.
\_ Alphas are evil now that they've sold out
to Compaq, Microsoft's puppet corporation.
\_ Perhaps Compaq is good now that it's
making Alphas, instead? |
| 1999/2/10-11 [Computer/HW/CPU, Computer/SW/OS] UID:15392 Activity:nil |
2/9 Any local mirrors of the wu-ftpd patches? (either src or x86 RPMs)
\_ what exactly was this exploit?
\_ no patches known for the academ-wuftpd yet. the exploit
was a buffer overflow involving really long pathnames
which can be created by an ftp user creating a very deep
directory strucutstructure in a world writeable directory such
as the incoming directory. |
| 1999/1/27-29 [Computer/HW/CPU] UID:15305 Activity:kinda low |
1/27 If I run a CPU intensive job with "nice", how much responce time
will I suffer, on a Linux box?
\_ 40 angtrom-meters. why don't you just try it? there are a lot of
factors. --aaron
\_ Depends on how much you nice it, doesn't it?
\_ The Linux scheduler is worse for this than something like the
lottery scheduler that runs on soda. But a CPU-bound job that
doesn't do complicated I/O may not be hurt too much by being
niced. It all _does_ depend on your hardware, what other
jobs are running, how much you nice it...
\_ response.
\_ I have a related question. What scheduling algorithm
does Linux use? Is it similar to round-robin with
multi-level feedback? |
| 1999/1/26-29 [Consumer/CellPhone, Computer/HW/CPU] UID:15299 Activity:high |
1/26 Suggestions on a DLT/70 drive?
\_ Aren't they all made by one manufacturer then OEM'ed? -ax
\_ yes, they're made by quantum. Buy one, is my suggestion. -tom
\_ Make sure you get one with a GOOD FAN, DLT7000's have a tendency
to run hot. If you like DLT, right now, DLT2000XT is best for
price/performance. DLT7000s have great capacity, and a great
technology in general, but are a bit more expensive. DLT7000s
also contain backwards compatibility to the very first DLT model
only included in DEC workstations of old, thus it can take a while
for a DLT to load a tape. Those MIT-wits made pretty good
hardware---DLT (sold to Quantum), DECetherworks (sold to Intel),
DEC Raid (Used to be the best you could get), Alpha (off to Compaq),
and PCs with quality fans---but, they sucked at marketing.
BTW, DEC had a deal with Samsung to let Samsung manufacture
Alphas just before DEC got bought out... I guess that deal's off now.
Too bad. -mtbb (sorry for the long rant)
\_ Samsung was and AFAIK is still making Alphas which are coming
out of an Intel fab.
\_ Samsung manufactures Alphas. Intel also fabs Alphas for
Compaq. --PeterM
\_ there are actually some pretty interesting rumors
about why intel and samsung are fabbing alphas. the
truth is definitely out there.
\_ Rumors? DEC wanted to get rid of Fab7. It wasn't
making money. So they used their patent lawsuit
to force Intel to take the fab off their hands, AND
forced a requirement that Intel fab alphas for them
as part of the settlement. The FTC made DEC
find a second-source for alpha, since Intel wields
near-monopoly power, hence Samsung's architecture
license. Samsung is hoping to make money, but I
think it's been a big loss for them so far.
Compaq seems to have carried on the DEC tradition
of having really stupid marketing, so perhaps
Alpha will be dead by 2001, unless COMPAQ does
something quick.
--PeterM
\_ Not quite. The hot rumor was Compaq had been
interested in DEC from a couple years back, but
did not want to pick up the Fabs also. DEC got
rid of the fab specifically to facilitate a sale
marketting, you'll still see alphas out there.
to Compaq. Microsoft engineered the whole thing,
to create a Compaq/Samsung/Alpha competitor for
Intel/Pentium.
\_ Dead? Hardly. By 2001, they'll be pushing systems
with 1+ ghz. If you've got the money, you can't
beat a DEC system in the unix world. (And no, it
doesn't run Linux). Even with their lame
marketing, you'll still see alphas out there. Its
always been a high end niche. It'll stay that way.
\_ DEC has a tradition of making fine, quality hardware.
Their PCs had fans which never wore down/broke, stuff I
am sure Compaq will yank out right away and replace with
the standard cheap crap you see on most PCs today. The
problem has always been DEC never knew how to market their
products. Take DLT drives for example, when they intro-
duced the DLT2000/4000 models a couple years ago, the
product was so good, they grabbed up HUGE portions of
Exabyte's market. Problem was, they under-estimated
demand for their product, and without a means to increase
production soon enough, they sold the technology to
Quantum. Quantum has made a killing off the DLT line.
If Samsung & Compaq would get their acts together, they
could really take chunks of UNIX market from other UNIX
and even NT vendors. But, Compaq's put themselves in a
situation similar to IBM's, where they have too much stuff
to focus on doing well in one market. Compaq could really
exploit the Alpha chip if they didn't have to worry about
what Bill Gates might think, etc. In our workplace, we
have 4 Alpha 8400 servers (4 GB memory, 4+ CPUs, 100+ GB
of RAID, 4 SCSI busses / per server). Each server can
run up to 9 different SAP Instances. For the SAME AMOUNT
of money, you can get an HP server with maybe 2 crap CPUs,
a SCSI bus which only recognizes other HP hardware, up to
1 GB of main memory, and enuff other resources to handle
up to 3 SAP Instances if we're lucky. HP doesn't even
give you a decent monitor - they expect you to buy X terms
from them for even more $$$. What do you expect from
$tanfurdders? Unfortunately, it's all in the hands of
Compaq's shareholders now. -mtbb |
| 1998/12/30-31 [Computer/HW/CPU] UID:15144 Activity:nil |
12/28 How does the AMD K6-2 chip compare to Pentium II, realistically?
\_ http://www.tomshardware.com http://www.anandtech.com http://www.shakeys.com (not sure
on the last one exactly but anandtech has a link to them).
\_ I hear it performs very well but you will have a harder time
trying to overclock it without nasty things happening.
Tperiod >= Tcko + Tp + Tsetup + Thold
-- former eecs150 student. |
| 1998/12/30-31 [Finance, Computer/HW/CPU] UID:15141 Activity:kinda low |
12/29 Does anyone know a hack around the free WordPerfect for Linux's
"This program will expire in xx days - please enter a license"
message? --cheapo poor guy who can't afford to pay $40 for a word
processor
\_ If you're an 3L33T HAQR D00D, you can use the 5UP3R S3KR1T C0DZ
page at <DEAD>livewire.corel.com/wp8LinuxReg/register.html<DEAD>
\_ Learn LaTeX - all the _really_ cool cs people use it.
\_ So latex in the cs world doesn't refer to the material used
in condoms but some nerdish program.
\_ That's more than enough reason NOT to use it.
\_ Uh, the license is free for personal use.
\_ yes, if you'd read the documentation, you'd find there's
a web page to go to to fill out a form and get a license
number. -tom
\_ Totally! Go get 'em boy! Sick'em! |
| 1998/11/19 [Computer/HW/CPU] UID:14982 Activity:high |
11/19 I don't get this. When Intel releases a chip with one bug in it the
government forces Intel to recall and replace all the buggy chips.
\_ uh, the "government" forced Intel to recall+reship?
\_ i think the government threatened them at one point. anyone
know for sure?
But when a company like Microsoft releases an OS with at least 50
bugs no one gives a damn. What the hell is up with that?
\_ at least 50? That's like saying, "Berkeley has at least 50
undergrads." Why don't you say "at least 20,000"? It might be
a more accurate estimate.
\_ When AT&T released an operating system that had 50 bugs, no
one gave a damn, and AT&T's stated policy was that they
*wouldn't* necessarily fix them... |
| 1998/10/11-14 [Computer/HW/CPU, Computer/SW/Unix] UID:14766 Activity:moderate |
10/11 What type of hardware does Soda run on? (Is this a FAQ? If so, where
should I be looking?)
\_ /csua/share/doc/machines
\_ wow, i never knew we had such awesome machines in the lounge.
last time i went there were a bunch of huge super-loud old
computers with monochrome monitors.
\_ they're still pretty loud, lots of background hum, but at
least they are color. --jon
\_ What's the deal with all the Ultrix machines? Is it possible
to go to Digital UNIX instead? --dim
\_ DU is not supported on the ultrix boxes' hardware..
\_ Would you like to donate 6 Alpha Workstations to the CSUA
so that we could run wonderful digital unix? --jon
\_ Aren't the DEC 5[05]00's Alphas? I'm not that up on
older DEC hardware. --dim
\_ DecSystem 5(5|0)00's run Dig Unix since they are
Alpha based. DecServer/Stations of the same number
run Ultrix and are MIPS based. Go Dec System Name
Recycling. --phillip
\_ For clarity: MIPS = ultrix, Alpha = DU.
\_ http://www.CSUA.Berkeley.EDU/computing/hardware/soda-mark-v.html |
| 1998/10/5-6 [Computer/HW/Memory, Computer/HW/CPU] UID:14734 Activity:insanely high |
10/5 overclocked my Celeron 300A to 450 MHz. ~jctwu/pub/mendocino.
\_ YouGo Girl
\_ Hey, and I put nitrous in my Yugo!
\_ the new Celerons are very well-engineered and -fab'd. My bet
is that Intel had to cripple them to 66 MHz bus speed so as
not to compete with the Pentium 2. -jctwu
\_ good for you. go away
\_ hmmm... well, i guess this is a step in the right direction
instead of outright deleting a post you don't agree with.
\_ good for you. go away
\_Intel Celeron sucks
\_ good for you. go away
\_ good for you. go away
\_ good for you. go away
\_ yes the L2 cache-less ones. read file for details.
\_ Even the zero cache ones easily overclock and are great
for gaminng where the loss of cache is usually over come
by the higher bus speeds. If you're concerned about poor
"Business Benchmarks", well... seriously, just how fast
do you want MSWord to wait for you to type something?
Nothing wrong with Celerons, old or new.
\_ is it possible that they overclock MORE EASILY, due
to no cache? (I'm CS, not EECS, and curious)
\_ yes, take 150. cache increases probibility of
setup time violations.
\_ no.
\_ yes (non-EE). 1) You don't get any traffic on the
system bus between the CPU and L2 cache => easier
\_ you never get traffic on the system bus between
the CPU and L2 anyway.
[ CPU ]
|
[ L2$ ] [ Mem ]
| |
===========system bus=============
overclocking. 2) The L2 cache is usu. off-die,
and thus not under a heat sink / fan. The cache
would get super hot. -jctwu
\_ hmm, right. What happened to my Hennessey?
Anyways, reason 2 still applies. -jctwu
\_ Merging different things together onto
a single package (ie proc + mem) usually
cuts down on power consumption. Check out
U.Wisconson Gallileo and Berkeley iram.
\_ Well, no dah. Probably 1/3 of the power
goes to i/o drivers. As drivers get
smaller to nonexistent, obviously power
consumption goes down.
\_old news. btw which motherboard are you using?
\_ abit BH6. read file.
\_ the BH6 rocks.
\_ what file? says "file not found"
\_ was there; updated (in pub dir, not ~).
\_ what's the whole url dodo?
\_ it's not a url you fool; "more ~jctwu/pub/mendocino"
\_ symbolic link now on <DEAD>.../~jctwu/docs/mendocino.txt<DEAD>
\_ So you liked hardware tweaking your celeron? Feel like an
Uber-hacker? Now try the dual-cpu-enable tweak.
<DEAD>kikumaru.w-w.ne.jp/pc/celeron/index_e.html<DEAD> |
| 1998/10/3-5 [Computer/HW/CPU] UID:14724 Activity:moderate |
10/2 Is there any advantage multitasking-wise of a system with N processors
over a system with one processor that runs N times as fast? (Of
course there are other factors like N-slow-processor systems can be
available before the N-times-fast processor itself is, and maybe N
slow chips are cheaper than one fast chip.) Thanks.
\_ MP's are worthless unless you have an OS that supports SMP or
something along those lines. Solaris supports it, Linux does
too but I don't know how well, FreeBSD will pretty soon, and
NT does but we all know how reliable that is. If you're using
Win95/98 don't bother with more than one proc. -jeff
\_ The more you scale a multiprocessor system the more bus traffic
you'll encounter so it will only scale up to a certain point.
This can be fixed with a larger cache on each processor but this
simply results in more cache incoherency causing even more bus
traffic.
\_ The fast processor can simulate the N chips in just about all
situations. Whereas N chips cannot simulate the single fast
chip in all situations.
On the other hand, if you have really really good I/O subsystems
if due to more
designed for multiprocessor,
and lots of multitasking, the N processor system can be a win.
Assuming they have good cache, etc.
\_ you must be this tall...
\_ this question depends on your workload. yes, there can be an
advantage due to saved context and keeping your TLBs. if the work
can't be parallelized you may end up losing on an MP. none of
this really matters unless you are doing serious crunching .-aaron
\_ it depends on why the single processor is N-times faster.
If due to more execution units, N procs gives you more regs,
if due to more registers, N procs gives you more exec units,
if due to higher cache bandwidth, etc. etc. There will
be situations where the N processors have the upper hand. -nick |
| 1998/9/28-29 [Computer/HW/CPU] UID:14689 Activity:moderate |
9/28 got any tips on buying a decent PC [Piece o Crud] piece by
piece for around $1k total?
\_ If you're in L.A., go to the Pomona computer show and buy pieces
there. http://www.pricewatch.com for prices. Buy high-quality hardware
at a good price point based on what you need or expect to need.
Celeron 300A and AMD K6-2 300 are good, cheap processors. Don't
cheap out on the motherboard. If you feel like it, you can mail
me or post a pointer to your desired set-up and I'll comment for
free. $900 (includes tax) is what I'm planning on spending this
weekend for a 300A w/17" CTX OEM monitor. -jctwu
\_ Celerons are worthless marketing trash, buy a K6-2. --dbushong
\_ 266 and 300s = sucky. 300A, 333 = good. Wait 1 week
and I'll tell you how the 300A is. -jctwu
\_ Uh oh.. someone's been looking at worthless business
"benchmark" numbers instead of http://www.tomshardware.com Silly
boy.
\_ The 300A's have double the onchip cache, for a total of 128k
of *onchip* l1 cache. Their performance isn't so bad. -PeterM
\_ Not only that, they are the most overclockable x86 CPU
ever... See http://www.tomshardware.com -mogul |
| 1998/9/21 [Computer/HW/CPU] UID:14637 Activity:high |
9/20 Anyone know where I can get a free linux based word processor?
Better yet, anyone know where I can get linux WordPerfect for
free? No, I will not use Word.
\_ http://www.linux.org/apps/applications.html
\_ TeX. it rules. -dpetrou |
| 1998/8/31-9/2 [Computer/HW/CPU] UID:14527 Activity:very high |
8/31 Anyone know how vector processors work? You'd think it would be
hard to exploit parallelism where you would issue one instruction
on several pieces of data. Isn't that wastefull of ALUs and cpu
space on general purpose instructions that don't require single
instructions on multiple data (sorry, haven't taken 152/252 yet).
\_ Vector processors are very efficient, since they allow you to do
loops in single instructions and suchlike, and are reasonably
easy to compile too, as well. Either talk to me if you want to
learn more, or look in the appendix of Hennesey & Patterson.
-nweaver
\_ for the loop for(i=0;i<32;i++) { A[i]++; } that would work
at the end of a vector operation!!) and low-overhead conditional
fine. But entire programs aren't like that and, thus,
cannot exploit vector parallism. Hence, on a normal
instruction you would use only 1 ALU and the other 31 won't
be used at all (is that how it works?). Why not just use
superscalar design instead since it works on all
of your sin of stupidity.
instructions?
\_ A lot larger sections of the program, especially those
inner loops in multimedia applications, can be vectorized.
And in general, you want a vector coprocessor on a
normal microprocessor. And a vector processor is small,
\_ obviously you are a moron. if by "every processor" you mean
"every processor in the intel x86 family," then you are obsolved
of your sin of stupidity. many number crunching processors (like
DSPs) have zero overhead loops or conditional instructions.
These just end up being not very useful ops on a general purpose
architecture that runs shit like windows and ms word.
\_ Vectorised SpellCheck!!!!!!!
\_ MasPar SpellCheck!
\_ Just curious, what do you use for a word processor, if not
MS Word or one of the similar apps from another company also
running on Windows?
\_ i use MS word and Emacs.
\_ I don't use word processors - plain text & html are all
the formatting I need, and vi/emacs do much better than
MS Word at editing either.
especially fixed point. A 32 bit, 8 lane (8 separate
pipelines), 2 functional units and one load/store unit
can fit in roughly the space taken up by a 16kB cache!
Vector processors are very economical to build because
they require almost no controll logic. -nweaver
\_ if you want to run word on a vector, well... get a clue?
\_ Isn't MMX intel's vector instruction set meant to
operate in desktop computers running word processors.
Okay, consider using vector inst to do scientific
computing. Exploiting parallelism would still be hard
because of the complexity of software. Many scientific
programs have multilevel loops and recursive procedure
calls and such nonlinear loops. There are also very
complicated data structures like gmp which primitive
cpu instructions could not handle (in case you were
trying to break and RSA key). What use are vector
ISA's if you can't exploit them on a general purpose
machine?
\_ You saying Dave Patterson is a twink?
\_ Dave Patterson is a twink.
\_ they allow low-to-zero cost synchronization (implicit synch.
at the end of a vector loop!!) and low-overhead conditional
execution (with condition masks); every processor should have
both of these, but none do. -nick |
| 1998/7/30 [Computer/HW/CPU] UID:14414 Activity:nil |
7/29 hmmm... http://www.idt.mdh.se/kpt/billsux.jpg \_ Hmm, intel design rules allow diagonal metal at a looser pitch. I wonder which metal layer this is? \_ Easy. This is a hoax, not an Intel processor. link:www.amazon.com/covers/0/69/102/125/0691021252.l.gif \_ Stop making fun of our President of the United States. \_ As opposed to my or somebody else's President of the United States? \_ Did Monica or Hillary work for Intel? |
| 1998/7/25-28 [Computer/HW/CPU] UID:14389 Activity:kinda low |
7/24 hello, does anyone knwo if there is a reasonable comparison of
perforance of x86 unixes? there is some circa 1996 stuff at
http://mosquitonet.stanford.edu/~laik/benchmarks but obviously
something more uptodate would be nice. this is for a high-end
machine, in case that makes a difference. --psb
\-jeeze louise: high-end is relative to it's class. by this i mean
a pentium 400 and not a pentium66 ... i dont know if one of the OSes
makes compromises to run nicely on lower end hardware. Obviously i am
not referring to a $24million dollar computer. --psb
\_ COMMODOR 64 RULES ALL
\_ "high-end x86" hmmmm
\_ as opposed to being on, say, a 386 ?
\_ "high-end Cray/Alpha/Solaris", okay... "high-end x86" hmmmmmmm
\_ Death to all the closed minded technology-at-any-price
bigots!
\_ what about a low end Cray? hmmm...
\_ I'll put a PII-400 against your EV4 Alpha or Sun Sparc
at any speeds they sell to the market. Wake up.
\_ Are you on drugs? While comparable in integer
performance, Intel chips still suck at floating
point -- and then you get to deal with crappy PC
hardware, too. --dim
\_ The last time you used FP was....? The rest of
hardware is whatever you put in the box. What
crappy hardware were you referring to exactly?
The same HDs made by seagate/wd/quantum, etc?
The same video chips/cards from asia? The same
memory from asia? Hell, even your mouse is
probably from the same factory in asia. No, I'm
not on drugs. I have my mind open and my
calculator on. I left my bible in the trash.
\_ Floating point is very important to many of
the applications I run. I don't even bother
looking at integer performance. PC hardware
is kludgy and lame. Remember that the next
time you're flashing your BIOS to get your
MMX-enhanced, PnP, jumperless, EIDE (ha!
although Sun has succumbed) controller
working. Intel makes good chips at excellent
prices. There's a lot more to a machine than
CPU performance though (e.g. bus speeds).
PCs are still mid-range and don't forget it.
--dim (who loves X86 boxes but isn't blinded
by Wintel)
\_ There's a lot more why the x86 architecture
sucks besides FP. FP is used a lot in lossy
decompression (like mp3) and high end
graphics.
\_ /dev/hme, zs0, ring overflow. Gogo SUN
cheap ass parts. --jon
(okay to be fair, even with these crappy
parts that have been around for ages and
have much better functional replacements
around now, Sun makes sparc workstations
work and work consistently)
\_ Er, um, a PII-400 will stomp an EV4 alpha on nearly
anything. EV4's are the same generation as Pentiums...
h0zer
EV5 (21164) alphas are comparable in int, faster in
fp. Ultras are also comparible in int to PII, and
also faster in FP. --PeterM
\_ Exactly my point. You're choosing to put the P2
against another chip based on integer performance
and then say it sucks against that chip by another
measure. Let's turn that around and compare an
Intel chip vs. an Alpha with the same FP and then
we can all say how much the Alpha sucks compared to
"a comparable" Intel chip.
\_ do you realize how stupid that is? Pitting Intel's
current fastest processor against a processor Dec
made two years ago, just because the processor Dec
made two years ago was as fast then as Intel's
fastest chip is now, is absurd. Match Intel's
fastest chip against Dec's. Or do a bang-for-buck.
\_ I dunno, our UE10000 is awfully zippy...
\_ Sparc, not Ultra. Just making the point that
not all sparc/alpha >>> all intel. Throw in some
price/performance numbers and intel starts looking
pretty damned good for a lot of things.
\_ Motorola 680x0 will always be here. You
are not ready for immortality.
\_ Apparently not. |
| 1998/6/9 [Computer/HW/Memory, Computer/HW/CPU] UID:14188 Activity:nil |
6/8 See the machine that is the legacy of the late Seymour Cray:
<DEAD>gopher.CCS.ORNL.GOV/SRC<DEAD>
6 Intel Xeon Processors, cray-like memory system, and FPGA-based
reconfigurable elements. --jon
- GEEKS!!
\_ Don't wanna read geek-talk? Get the hell out then.
\_ Da CSUA baby! Love it or leave it. - tpc
\_ how aobut you leave instead? |
| 1998/5/8 [Computer/SW/OS/OsX, Computer/HW/CPU] UID:14073 Activity:nil |
5/7 Motorola's High-Performance Vector Parallel Processing
Expansion to the PowerPC Architecture
http://www.mot.com/SPS/PowerPC/AltiVec/facts.html --jon
\_ Is this really going to help, if the CPU can't stream data
in and out fast enough? I thought existing CPU's all had
trouble keeping their units fed, let alone a vector unit. --PeterM
\_ Real vector units (which this is not) are much more efficient
at keeping functional units fed for a given memory bandwidth,
as a vector processor can pipeline memory much more efficiently
than a cache based machine. Unfortunately, the vector leinght
on this proposed ISA is WAY WAY WAY too small for this to work.
\_ I think that's why it only exists on backside-cache CPUs?
\_ It'll probably help as much as MMX. -tom
\_ And that's the important thing, isn't it? Because now
future Macs and Mac software can have an "Now optimized
for AltiVec(tm)!" sticker on the outside of them,
thereby assuring sticker and I-dont-know-what-it-means-
but-damn-I-the-consumer-sure-need-it parity with Intel.
Macintosh forever!
\_ and cdaveb can say, "Oh _yeah_? But yer PC's don't
have _AltiVec_!! Neener-neener!"
\_ If you want to read more, MacKido has an article:
http://www.MacKiDo.com/Dojo/AltiVecVsMMX.html |
| 1998/4/6 [Computer/HW/CPU] UID:13904 Activity:nil |
4/3 Does Linux run on any other platform besides PC? Thanks.
\_ Alpha, PowerPC, Sparc
\_ Also ARM, 68k, MIPS, although the above are the only ones that
are really usable. Linux on the PowerPC is getting quite
mature. If you want Unix on a non-PPC Mac, there is also NetBSD. |
| 1998/3/25-26 [Computer/HW/CPU] UID:13861 Activity:insanely high |
3/25 What is the human facination with the letter X? We have:
DX (basic), LX (luxurious), EX (extra luxurious), FX (special effect),
GX (Nissan GXE), MX (Cyrix MX chip), RX (Acura/Integra), SX (486 SX),
TX (Fujitsu TX), XXX, X-files??? Why???
\_ Of the 25486 words in my machine's /usr/dict/words file, 636 of
them have the letter "x". If there were an even distribution of
letters among all words, then ~980 should have an "x". At roughly
2/3rds the purely staticially expected word count, I don't think
there are as many "x"'s in use as there should be and advertisers
recognise this and try to make up for it a bit by throwing in a few
extra "x"'s here and there. The letter is woefully under utilised.
Use an X today!
-President and Co-founder of the Society for the Better Use of X.
\_ Well, exxxxxxxxxxcuse me!
\_ XXX-Files!
\_ Dont' forget X window system...
\_ and the worst gaming API ever - DirectX!
\_ XXXXXX HOT NAKED XXX BABES IN POPCORN BUTTER --- ONLINE!!!1!! XXXXXX
\_ Cuz X is 'exotic' -- it seems special in marketing drool-speak
\_ "sex" ends with an X, so everything else wants to.
\_ The letter X has special appeal to the average clueless consumer
who never really know what he/she really wants in life. It's all
a result of marketting research. |
| 1998/3/16 [Computer/HW/CPU] UID:49813 Activity:nil |
3/16 Does anyone know where to obtain information on PC serial port
interface, for input AND output? Thanks.
\_ Pins 2,3 for send/receive. What more do you need?
\_ yeah, knowing the port #'s and irq's isn't very
useful.
\_ IRQ 3 and 4, 2e8, 3e8, 2f8,3f8. Anything else? |
| 1998/3/9-12 [Computer/HW/CPU, Computer/HW/Display] UID:13783 Activity:high |
3/9 Anyone have experience using a computer video board/computer
monitor as a replacement for a TV? What video boards do the
trick? Are any of them supported by XFree86? --PeterM
\_ if by supported you mean simply that there is an xfree86
server for the video card, then the ati cards are supported,
at least by the latest release. the tv quality is quite
good too.
\_wHAT IS XFree86? -too ignorant to sign his post
\_ It's an X based client free of all x86 instructions.
\_ It's a "free" version of the letter X. The Gnomes of Zurich, who
hold all rights to the use of the alphabet, in any form, were
dismayed at the incredibly low licensing revenues brought in by
"X" in comparison to more popular letters like "E". In an effort
to stimulate interest in the letter X and increase its usage, they
released a public-domain, royalty-free version a couple of years
ago. However, since they didn't want current customers to feel
shortchanged, they released an older version, designed under
contract by the American Type Foundry (now suspiciously out of
business) and first released in 1886. Hence the "86".
\_ And you wonder why quality fonts cost so much?
\_ While it's true that the Gnomes of Zurich (ZOG)
hold all rights to the use of the alphabet, in
any form, their license is set to expire this
May, when the IANA will hold conference commitee
hearings to determine world readable font
ownership rights.
\_ That's what you think. The Gnomes have clearly stated
that if they're challenged on font-ownership rights,
they'll stop licensing even their most basic patents
(such as the ones for water and gravity). Is guaranteed
access to a "2500 fonts for $3.99" CD worth spending the
rest of your life floating around, thirsty? -- kahogan
You are clearly misinformed
about the 86 part, as the American Type Foundry
people dropped the ball when they designed the
1786 version to be compatible only w/ manufactured
wood pulp products, ignoring almost entirely the
burgeoning digital font transmission and display
market which ZOG was better prepared for. There
were no "suspicious" circumstances at all, and
I would advise whoever wrote this unsigned
missive that you check your facts before
misinforming the public so blatantly. -tpc
\_ No. http://www.xfree86.org
\_ http://www.xfree86.org is currently the subject of
a trademark infringement lawsuit from GOZ.
GOZ's trademark, of course, is being challenged
in court by ZOG, which claims to have nothing
to do with http://xfree86.org. GOZ has countersued,
claiming that ZOG is deliberately trying to
dilute the TM. The software supposedly available
at xfree86 is, as far as can be determined,
a complete fabrication.
\_ Dammit, so _that's_ why "xf86config" never seemed
to work right with my hardware . . . |
| 1998/3/4 [Computer/HW/CPU] UID:13758 Activity:nil 57%like:13756 |
3/3 Intel is cool in my book now. They adopted StrongARM.
\_ they're killing StorngARM, by jacking up licensing or somesuch,
I hear.
Well, Merced looks pretty good too. -muchandr
\_ And you told me not to work there ;) -emarkp
\_ I am not saying it is a good place to work. They are cool
for finally recognizing that x86 must die. -muchandr
\_ I think they've been aware the x86 needed to die a long time ago
but market forces prevailed over technology choices. You can't
just up and change your basic cpu architecture over night and
expect everyone will just recompile their OS, apps, reinstall,
buy new chips, all new whatever, etc, etc, just because you made
a poor long term architectural decision as a company 15+ years
earlier.
\_ Apple, PowerPC. Whatever you think of the platform, the
transition itself was comfortable and smooth (simile
self-censored) -pld
\_ Much like when I make love. -anonymous
\_ It was reasonably smooth. There are still problems running
old software on the new platform. It wasn't perfect. That
means lost money to some companies (customers) who would
see the hassle but not the benefit (I don't entirely agree
with this attitude but its the way things are). Thus,
squeezing every last drop out of the x86 was a better move
for Intel than trying to switch to something else (what?)
and do emulation/translation. Maybe the merced or whatever
will be the cool next step and whatever comes after won't
do x86 at all. But given the incredibly deep level of
x86 entrenchment, I doubt you'll see an x86 free world in
less than 10 to 15 years. The current line of chips from
the 486 to P][ are powerful enough for what the bulk of
business users need a computer for. They're not going to
switch and upgrade everything just because a new version
of something is out there. That costs money. |
| 1998/3/4 [Computer/HW/CPU] UID:13756 Activity:nil 57%like:13758 |
3/3 Intel is cool in my book now. They adopted StrongARM.
Well, Merced looks pretty good too. -muchandr
\_ And you told me not to work there ;) -emarkp |
| 1998/2/18-20 [Computer/HW/CPU] UID:13691 Activity:high |
02/20 INTEL PENTIUM 200Mhz $115
INTEL PENTIUM 166mHZ $89
INTEL PENTIUM 150mHZ $75
INTEL PENTIUM 133mHZ $55
If interested, please email swings
\_ Are these old processors you have, or is this a business?
\_ well, they're definitely old processors....
\_ okay, you got me there. But I meant, is he selling
'pre-owned' or not. |
| 1998/1/27-28 [Computer/HW/CPU] UID:13568 Activity:high |
1/26 Compaq has bought Digital Corp. for 9.6 billion dollars check out
http://www.news.com/News/Item/0,4,18440,00.html
\_ I don't get it. Compaq is a very low tech, system component buyer/
builder. DEC is a very high tech, leading edge company. What do the
two have anything in common??!?
\_ Nothing. Now the merged company has both ends. Got it?
\_ FUCK ALL THESE REPLIES. COMPAQ WANTED ALTAVISTA
\_ COMPAQ WANTED ALTAVISTA
\_ $9.6 Bil for altavista?
\_ Dec... 'very' high tech? Not in the last 5 years. They
had to settle for NT when their market collapsed. Ever since
the alpha died a horrible market recption Digital has been in
free fall from high-tech to no-tech. Compaq got a raw deal.
\_ Mabye, maybe not. DEC was more than "the alpha company".
Time will tell of course.
\_ DEC is still a 'high tech' company. No one's beaten their
CPU, they've only matched it. They've got other
technological shortcomings, but they're leading edge.--PeterM
\_ Compaq's money?
\_ The two have very little in common. It's a good match for
Compaq, tho, as C's been trying to move into the higher end
of the enterprise / server market... with Digital's sales
and finance arms, Compaq might end up doing quite well.
\_ Bull will crush you all. -Evil Euro
\_ okay, DEC sucks, CPQ sucks. Other irrelevant stuff erased. |
| 1998/1/21 [Computer/HW/CPU] UID:13524 Activity:very high |
1/20 So, this Intel compression proxy thing... is this similar to that
proxy called "Transend" that some Berkeley students came up with?
\_ sorta, but not as scalable and they stole the ideas from us.
(The transend group gave a talk to intel about a year before
the started working on the project)
\_What the $%$! for? An open presentation, fine. But specifically
going to present things to a company that is known to compete in
things like that? yik.
\_ Umm...this is a University - we believe in sharing
information - we're not in competition with companies.
(If we were competing with Intel, you think they would
be donating millions of dollars worth of PC's every year?)
\_ acutally my bitch isn't that they made it. But
that there is NO credit given to our ideas and that
Intel is trying to patent a lot of things we had done
earlier.
\_ If you published, they can't patent. You *did*
publish, right? Oh? You didn't? Then how do we
know you did anything at all? How do we know *you*
didn't steal *their* ideas?
\_ uh we published, lots. The point is now we
need to defend our work and Intel is being
lame, espeically considering they KNEW that
they were taking ideas from glomop,
\_ Maybe you should give it up and go work for
Intel on their new Transcend project? I hear
they may have some lower level positions for
some bright young newbies looking for a start
in the world. If you're lucky, after a year
or two of doing backups, they might let you
write a few small snipets of code.
\_ wasn't there a URL for it somewhere locally? |
| 1997/12/3 [Computer/HW/CPU] UID:32172 Activity:nil |
12/2 Pentium II sucks big time!!!!!!!!!!!!!!!!!!!!!!!!!!!!
\_ Explain
\_ It's a tad expensive but if you can't afford it you don't need
it in the first place. Buy an AMD or Cyrix or something. |
| 1997/2/10 [Computer/HW/CPU] UID:32061 Activity:nil |
2/10 Has anyone ever tried rejumpering a Pentium 75Mhz to 100Mhz?
I've tried rejumpering it to 90Mhz and it works great. Will
raising it to 100Mhz fry the cpu?
\_ Can you overdrive a 100Mhz as well?
\_ No. _ALL_ Pentiums are actually the _SAME_ chip!!!1!!
It's just an Intel/Microsoft/government conspiracy to defraud
the consumer by fooling them into thinking otherwise.
D0N"T B3 F00L3D!!!1!!!!!
\_ Bzzzzt. The "Pentium" is not the same chip as the
P54C or the P55C. You can probably push your 75MHz
chip to 100MHz. It may freeze up, but won't hurt
the chip. Just put it back to 90MHz. Try this page:
http://sysdoc.pair.com and its overclocking guide.
This chip will never go faster than 100Mhz tho.
--nevman
\_ Intel did mark a lot of better qualified chips as 75mhz to
have a cheaper price point...so you might not be overclocking
to do it. If you experience lockups, then you're out of luck.
\_Can you elaborate? what's the difference between
cpu's with different clock speeds?
\_ The chip failed faster tests, so it's "marked down".
\_ If you see smoke coming out of the chip, then you know that you've
overclocked it too much.
\_ Is it safe to overclock a 90MHz Pentium? If so, how would
I go about doing it?
\_ You need clue.
\_ that's why I asked the question, twink. |
| 1996/10/19 [Computer/HW/CPU] UID:31948 Activity:nil |
10/18 Monday Exponential Systems releases 533MHz PPC chip. The Intel,
The! http://www.news.com -dbushong
\_ Huh Intel Huh?
\_ Watch Simpsons! Have you no culture?
\_ Bablyon 5 is true culture - the Narns & gospel scene of
this week's episode (you know what I'm talking about if
you saw it - if not watch the Sunday rerun) is art of
the highest degree.
\_ Essence of Clockwork Orange, distilled into 120 seconds
of tv. WOW! |
| 1994/2/14 [Computer/HW/CPU] UID:31484 Activity:nil |
2/13 Want a free Pentium? Call 1-800-955-5599, and ask for 'info pack #90'
Intel will send you free info on their Pentium processor, including a
keychain with a Pentium that failed the burn-in testing. |
| 1993/11/18 [Computer/Companies/Apple, Computer/HW/CPU] UID:31431 Activity:nil |
11/18 In his inaugural Comdex address, Apple Chairman Michael Spindler
promised a RISC (reduced instruction-set computer) machine on every
desk, and a Newton for every pocket. |
| 5/16 |