Motherboards with i845 Chipsets

manplusdog writes: "Dan reviews a couple of i845 motherboards here and lets just say he doesn't hold back! "Mainboards For The Stupid" is the verdict. I have no affiliation with Dan or his site (aside from being Australian) but found the review..... entertaining. Cheers"

Comments

slashdotted already

[AntiNorm]

from the penny-pincher dept.

Well, now that we have successfully Slashdotted his site...

Re:slashdotted already

[sasha328]

It doesn't look like it is slashdotted; I can get to the site, however, the link seems to be to a non-existant page. I also did a "google" search on that site for anything 845, and nothing was found. Also nothing about i845 in the reviews section.
What's up, anyone?

Re:slashdotted already

[ragnarok]

Yeah, like we've never heard that around here before.

Good article

[Inthewire]

He reinforced my belief in the value of a competitive market...without AMD, Intel would have a lock on all of our business. It's interesting how destructive human nature can be...logically, the thing to do is run a single company that can leverage suppliers, research, manufacturing, distribution, administration, etc. This would reduce all the redundancies in the market and allow for superior products at reasonable prices.
Advertising could be focused on actual products, not competitive differentation. If something new was developed by this company, they would only need make the value known...no more blue men.
INstead of this utopia, when a single company gains the majority of the market they tend to maximize profit instead of customer value.
It's a hell of a world, isn't it?

Re:Good article

[Olivier Galibert]

They don't really have a choice, it's the law. It's called maximizing shareholder value. Enjoy.

OG.

Re:Good article

[gazbo]

Yeah, hopefully this will kick Intel into working a bit harder for their money. I remember the day when I wouldn't wish AMD on anybody - they were scarcely better than Cyrix (actually, that's too nasty to say about anything) but really, they were just making cheap, plastic (metaphorically) copies of Intel chips that ran slower, but were dirt cheap. Then the Athlon came out, and I had to take everything back, as they made a better chip for less money. Strange, you'd have thought Intel would be more concerned.

Hopefully with the ludicrous prices for the high MHz P4s that we need to even compete with an Athlon that is a fracton of the price, people will stop licking the TV when Intel adverts come on, and instead buy the obvious choice. Maybe then Intel will have to come up with a decent chip, rather than just squeezing out an extra 200MHz every few months for a mere £500 extra.

I have a P3 500, btw. Bought it just before 1GHz Athlons came out for about the same price. Doh!

Re:Good article

[CLorox]

Oh yeah, but they came out with a great chip before the Athlon, remember the 5x86! :) 133mhz 486 class chip that could actually turn an ailing half dead computer into something that could play quake full screen / play mp3s. You could even overclock it to around 160mhz!!! I ran two of those chips for years before the CPU fan died and it burned to a crisp while I was on vacation.

-Adam

Speech output

[kingdon]

The coolest part was the ASUS board which speaks for power on self-test errors, rather than the age-old cryptic beep system. And the fact that you could download new messages. Anyone done the Klingon translation yet?

Re:Speech output

[espresso_now]

My AOpen AK-73 Pro(A) also speeks diagnostic information if something isn't right. It'll talk in 3 or 4 different languages too. But that's not the best feature, the PCB is black... :)

Re:Speech output

[Doomdark]

As long as it can be muted before I ever hear a single syllable, it's fine with me. :-)

I guess I would pay (if I had to) for a PC that doesn't talk back to me. Hmmh... but think about this; if Windows tried to soothe you when BSOD arrives, what would you do? Feel relaxed and smile?

My preference for output method might be a simple led-display (with just one or two digits, referring to error codes), but I'm not sure how that could easily be pre-configured to be fail-safe like simple aural devices (as they need not be connected to case). Or perhaps a set of red/green leds, indicating progress of self-diagnostics (although, once again, to see the leds, they'd need to connect to the case or you have to open the case).

His page is upside down

[Anonymous Coward]

This Australian forgot to add the hemisphere qualifier to his HTML tag. When will they learn?

<html hemisphere="south">

Re:His page is upside down

[Chagrin]

I'll bet that once the server gets slashdotted it will go down counter-clockwise.

Re:His page is upside down

[Morbid Curiosity]

<html hemisphere="south">

You might want to try <html lang="EN-AU-STRINE">, too. But that's only for those who're as pedantic as a tree full of galahs, mate.

Custom Errors

[Myriad]

Instead of using lights or mere beeps to tell you about its Power On Self Test progress, it speaks its errors,in a quite comprehensible female voice.

The P4B also comes with a Windows utility that lets you convert WAV files to make your own error messages.

In related news, Asus will begin shipping the Custom Error Pack with errors including:
- I'm sorry Dave, I'm afraid I can't do that!
- It's Microsofts Fault - Really!
- BSOD my ass!
- Doh!
- Need Beer!
- I've been Slashdotted!
and many more!

Re:Custom Errors

[samason]

I'm taking bets for the amount of time someone will take to record a standard set of POST beeps. . .

• 1 year
• 1 month
• 1 week
• tomorrow
• its already been done

The Pentium 4 is worth the extra price.

[Starship Trooper]

Now, Athlons may be fine for gaming and other unsubstantial desktop usage. But for the workout my Fortune 500 company gives their computers, AMD's products cannot compare to Intel for one simple reason - Reliability. This is where the value of the Pentium 4 comes in. In my experience, Athlon processors, while maybe a few percent faster than their Intel counterparts, die at least three times more frequently. When you have to buy the same AMD processor thrice as often as the comparable Intel one, the cost difference becomes negligible.

There are a number of other reasons the Pentium 4 platform is a better value than the Athlon:

• Rambus memory. Despite the common anti-Rambus sentiment here at Slashdot, RDRAM is of consistently higher quality and better performance than SDRAM, especially in quad-processor situations where memory bandwidth is everything and even DDR SDRAM becomes a bottleneck. Not so with RDRAM.
• SSE support. SSE2-enabled code beats the pants off the Athlon in performance. My company's heavy data processing algorithms depend heavily on SSE2 optimizations in order to process gigabytes of data in real time.
• Commitment to open source. Linux serves an important rôle in our work, and it's good to know that Intel has continued to support Linux and open source, and that part of our purchases goes to fund that support. Meanwhile, AMD jumps on the "XP" bandwagon with their new, specially-designed-for-Windows chips. That sort of behavior is detrimental to both our business and the entire tech economy.

So, I urge you all to transcend the Slashbot stereotype and realize that Intel truly provides the best value for business and home use. Just because they are the big player in the market don't make them bad.

Re:The Pentium 4 is worth the extra price.

[startled]

I have taken it upon myself to create an account, as has been suggested. As for my troll, I am surprised by the number of people duped by it; this is easier than I thought. Especially since I inserted numerous clues that it was all bogus ("O(n) NP-complete" indeed) for the more clueful people to see, as is proper troll protocol (at least, it seems it should be proper protocol to me). Now, alas, I just need to wait for inspiration to hit for my next attack.

Inspiration didn't take very long to strike, I see.

Re:The Pentium 4 is worth the extra price.

[at_18]

???

P4 and Rambus memory on quad-processor situations: I'm not aware of any P4 multi-processor motherboard

SSE2 for gigabytes of data in real-time: you must have unbelievable high speed hard disks in multiple RAID setups, if you can process realtime gigs of data with any processor

Open source: I won't comment on this one, since I'm not informed enough.

Well, troll score 2 out of 3.

Re:The Pentium 4 is worth the extra price.

[Watts]

I personally have had no reliability issues with Athlon systems, nor have any of my associates (with the exception of one faulty motherboard). While this personal point isn't going to prove anything for overall reliability worldwide, it makes me think when I see half a dozen systems running night and day in constant use with no crashes, freezing, or hardware failure.

Let's address your issues:

• Rambus memory: I'm not against Rambus. Maybe on an intellectual-property standpoint, or on a corporate standpoint, but they make some good memory. However, Intel has done an amazingly mediocre job of taking advantage of such memory in their chipset offerings. They obviously aren't going to get the performance gains that have been seen in game consoles because they don't have a unified memory architecture, but the fact remains that Intel engineers have had difficulty pulling the possible performance out of Rambus memory.
• SSE Support: As you've stated, SSE2 code does some really nice tricks. For "heavy data processing algorithms" it doesn't really have any competitors in the x86 world, yet. However, this is really limitjng the scope of applications, as not every program is going to be able to take advantage of this functionality. In fact, most won't. Overall, SSE2 is nice but takes some attention to optimize for. Whether a lot of mass usage programs will take advantage of it is yet to be seen. I'm not going to say anything too negative about it, because it is something that can be used well.
• Commitment to open source: Amazingly poor naming schemes aside, I don't believe AMD has any less of a commitment to open source than Intel does. I have a friend who was employed at AMD over a year ago who was paid to optimize software like glibc and gcc to take advantage of the Athlon processor. AMD's x86-64 has public specs and x86-64.org is hosted by AMD to showcase ports of open source projects to this new processor.

The Pentium 4 is a useful platform, but there are viable alternatives as well. Just because one piece of technology is good does not mean that others are bad. I personally would gladly use any stable, well-performing system that fits the given task.

Re:The Pentium 4 is worth the extra price.

[spiro_killglance]

"SSE Support: As you've stated, SSE2 code does some really nice tricks. For heavy data processing algorithms"

N.B. Note SSE2 code only applies to Double
precession floating point code.

For single precession SSE/3D Now to the same
jobs. Need Quad precession, your in software
emulation and its real slow.

Despite SSE2, the Athlon still rules at ScienceMark

Intels SSE2 autovectorising compiler still has
a lot of issues for general use.

actually i am kind of bummed.

[motherhead]

All my personal boxes are currently Athlon 1.2 and 1.4 GHz 266FSBs running DDR, except for a little sad rarely used mother that was once my main box, a 1GHz P-III on a Asus "Black Pearl" BX board.

So I love AMDs, they are swell. But there is one thing about Athlons that frosts my ass, well no, the opposite. I have had to build in the odd year of 2001, twenty-two separate AMD Athlon/Thunderbird boxes. I have had seven Athlons burn on and die on boot up (stinky silicon).

I am not a retard. And that is just unacceptable.

I have never dealt with a chip as volatile as the Thunderbird. Some are just hardy little bastardos, others need a level of anal retentiveness that borders on owning ones own clean room. For me and my absolute need to have a box that makes apps open before I can remove finger from the enter key, or off the mouse on the second click. This is okay. When I am building a blah beige business box, for a client, or a friend or Auntie Ann. Then this makes me borderline homicidal.

The fact of the matter is, any monkey with a hammer can knock together a P-III box. Intel chips tend to be as robust as those freaky bubble glass ashtrays that weigh fifty pounds. I can knock together a P-III box and have an operating system installed in an hour, mostly while I am doing something more important then watching Win2K load or whatever.

I honestly wanted to see a nice Asus/Abit P4 board available so I can do more of the same for business clients ("Oh! goodness Bob! look! ONE POINT EIGHT GIGAHERTZ!!! INTEL BOB!!! HAVE AT IT BAYBE!!!... But first, pay me.")

Cheap boxes that run as stable and reliable as hell and can be assembled almost by remote control rock, the extra cash keeps me in Geforce 3 cards and klipsch speakers and other shiny things I see in the forest. I would be happy as a clam to see this whole i845 thing straighten it's wings and head into the promised land that the BX chipset promised us exists. Speaking of BX, that Asus black pearl box in the corner. It's not nearly as fast as my other three Athlon boxes, but damn it, it is as reliable as my subzero fridge.

As for myself, I will stick to my yummy AMD goodness until the data becomes more compelling otherwise. I am still a sucker when I notice that something is really "noticeably faster"

Re:actually i am kind of bummed.

[Barbarian]

Golden orbs suck, they're worse than a Volcano II, which is barely adequate.

Sounds like the clips suck on those orbs, or you're putting them on rotated 180 degrees the wrong way, causing a miniscule air gap on the core. You know, you can put the heatsink on when the board is out of the case, and then look carefully to see if it's on properly. Also use Arctic Silver II thermal compound..

Re:actually i am kind of bummed.

[Skuld-Chan]

Thats really odd - since I've been building lots of AMD K7 based systems as well - I've never had one die on me - and I've never had a bad part. I don't know how many i've built, but its quite a lot.

I have however had people bring systems to me that didn't work - just two - both had bad procs, not because they were bad from the factory, but because they either a) used thermatake cpu coolers (please - don't use those) or b) put the heatsink the wrong way around - both of which will crunch the chip - which is relatively fragile being a ceramic package. Yeah - I know you can use thermatake coolers, but in my experience 90% of everyone who buys them doesn't seem to know how to properly use them - I don't recomend them myself - not only are they not the most efficent, but they also cost more.

So - you might want to switch heat sink manufacturers or check your supplier for defective parts.

Re:actually i am kind of bummed.

[Alcimedes]

lol, reading how many athlon cpu's you've fried, i've got to wonder if you aren't somehow screwing up the heatsinks. in light of last week's article here on athlon cpu's burning up when heatsinks aren't attached, while the intel's would keep on trucking.

wonder how many of your intel superboxes aren't running at half speed 24/7 to keep cool. he he he.

Re:actually i am kind of bummed.

[motherhead]

I had been aware of that actually, though you win the prize for bringing that to light first. No, that is not what killed the poor bastards.

and i like the golden orbs, though i have seen better since i picked up that lot. still the boxes that i have used them on stay quite cool even after doing a quake III demo break in. (though the nvidia cards get hot, have to bring air over those bastards)

damn i need sleep.

Re:actually i am kind of bummed.

[dasunt]

Building 1+ GHz Athlon-based systems for a living, I am familiar with a large amount of athlon based systems on several motherboards, (computer grunt gives one interesting knowledge).

Now even though we use the default OEM heatsink, which sucks, I've never seen an Athlon system with a properly installed processor overheat. This includes systems that are in older, forced air-heated homes with several pets (ewww, cat-hair is evil). The OEM heatsink is probably one of the poorer heatsinks, an Athlon under heavy load can reach 60C, and at idle on an operating system like win98, it doesn't even reach a low of 50C. Win2k and linux systems do better on idle, because of the HALT instructions sent to the CPU by the OS.

Please note that the temperature measurements were done using the the sensor on the motherboard. And 60C is still 35C cooler then the top limit Athlon sets for their TBirds.

Re:actually i am kind of bummed.

[ZeroConcept]

Who are you? Edward Scissorhands?

Re:actually i am kind of bummed.

[Mr Fodder]

I am not a retard. And that is just unacceptable.

That's gotta be the oddest statement I've read in quite some time.

Re:actually i am kind of bummed.

[motherhead]

okay now that is funny.

for the other guy, a nice haiku for you mister English major.

I'm not a retard
unacceptable came next
haikus are for fags.

thank you very much ladies and gentleman and goodnight.

I agree.

[Joe Groff]

While we aren't Fortune 500 where I work, we gave Athlons a try and, sorry to say, they just didn't make the cut. Too many hardware issues and not enough performance to justify putting up with it. Too bad the moderators will mod you down for telling the truth =(

Re:I agree.

[Listen Up]

Actually, I don't usually agree with too much of anything I hear on Slashdot (although I do definitely agree on some people's points here and there) but I don't understand you at all. The main problem with AMD and their associated hardware is that Windows 9x/ME/2K does not come with any decent support out of the box. That would be your hardware issues. Also buying "cheap as hell" motherboards will also lead to your hardware problems. I build computers for a business I have been running for years. I build AMD and Intel. AMD chips "burn up" only when you do not have proper cooling in your case and try to do stupid things with your chipsets. I have a Pentium III 500 I am typing this on with one huge ass fan and huge ass fan cover enclosure over the chip and chip housing. It looks and works very, very well. It basically looks and works like a steam hood over a donut fryer, if you have ever seen one of those. That is how it came from the manufacturer. I have never had a problem. As well as the AMD chips which I have never had a problem as well as I have proper ventilation over the chips. You buy a $40 motherboard, an AMD Athlon, and a $10 fan you are going to have problems my friend.
Advice: Fire your Tech guy. Who is a f*cking retard and is most likely soaking your company for more than 3 of him/her are worth. I am sorry you have had such problems. Competence would have not let any of that happen.

You bastards

[Anonymous DWord]

I wanted to read that! Ah well...

Repeated attempts failed to load this page completely. There may be a problem on the server.

Yeah, I'll say. Maybe if I hit 'reload' every 5 seconds, I'll get through faster. :-P

Also at Ace's

[kinnunen]

As always, Ace's has done a very good and analytical review of the new hardware. i845: SDRAM and the Pentium 4

I also recommend checking out their new PC1066 RDRAM review, which really shows shows you just how bandwidth dependant the Pentium 4 is.

Re:Yawn..

[Arker]

Why would any self respecting slashdot reader not opt for a DDR Athlon??

Well there's a couple reasons I can think of. First off you might be looking to upgrade an Intel board you got for free *cough* but in that case you wouldn't be buying one of these motherboards. :)

The second reason would be thermal protection. Intel build a little thermometre into their chips, along with some circuitry that'll turn the sucker off in case the temperature goes way over where it should be. Which isn't such a huge thing, if you use proper cooling it shouldn't matter, but in some cases it's probably worth thinking about.

If you do get an Athlon, be sure and cool it properly. They'll keep processing till they burst into flames... :)

DOH! Wrong link

[kinnunen]

Here is the real link to the PC1066 review: http://www.aceshardware.com/Spades/read.php?articl e_id=45000217

Re:Yawn..

[Anonymous DWord]

If you do get an Athlon, be sure and cool it properly. They'll keep processing till they burst into flames... :)

Which is less than a second without a good heat sink, apparently... :) Tom's Hardware lit some on fire for a test. I'm glad he does shit like that, so I don't have to.

Athlon Cooling

[Greyfox]

My cheap ass no-name Athlon motherboard comes with a handy array of temperature sensors and a bios-settable emergency shut-down feature. If I feel like putting some work into it, I can also access the sensors from Linux and initiate a shut-down if things get too hot, which they will only ever do if I lose one or more cooling fans. Logic to do this on the processor iself is, in my view, completely redundant.

Re: Athlon Cooling

[Inthewire]

Yeah? Wrong, unless you use some pretty aggressive settings. The time it takes software to detect the problem is 5 times longer than it takes an Athlon to self destruct.
Tom's reported a rise of a few hundred degrees per second...no software solution will catch it, unless you've found or written one in the last three days.
That said, I still choose AMD over Intel...I know how to keep the cooler locked on and finctioning.

Re:Athlon Cooling

[crimoid]

Given that the P4 will just throttle back to cool off (thereby staying up and operational with no data loss) rather than requiring a motherboard manufacturer to build in extra functionality to only _shut_down_ the system IMHO it seems that having the processor control this function is the ideal solution.

Re:Athlon Cooling

[GroovBird]

If you cared to read the article on Tom's Hardware, then you might think again. All those nifty sensors work shit when your TBird fires up to 300C in just a few seconds when your cooler falls off!

Dave

Re:Athlon Cooling

[Animats]

Actually, AMD finally has put thermal sensing on-chip. The AMD Athlon 4 model 6 has a thermal diode on-chip.

Re: Athlon Cooling

[tauntalum]

Yeah? Wrong, unless you use some pretty aggressive settings. The time it takes software to detect the problem is 5 times longer than it takes an Athlon to self destruct.

Tom's Hardware showed the circumstance of having no heatsink or fan. The more likely case of losing a fan would show a more gradual rise, especially if there were at least one auxiliary fan.

Reviews are cool, but whats the best hardware now?

[BrookHarty]

Whats the fastest Intel motherboard for P4s? Ram?
Whats the fastest Amd motherboard (Via chipset?)

I can find great prices via www.pricewatch.com But where can I find the best motherboard? I like asus, but which one is the fastest for intel and amd?

I read sites like toms hardware, sharkys, via harware, extra, but if I want to build the best, where is a good place for fastest hardware out NOW that I can find on pricewatch?

Tom's Hardware has a review too

[Anonymous DWord]

Here: http://www6.tomshardware.com/mainboard/01q3/010702 /index.html

The question now is, who will be interested in it? It is true that it will make Pentium 4 much more affordable due to its PC133 SDRAM support, but its lackluster memory performance impacts Pentium 4 so badly, that it makes AMD's Athlon an even more attractive solution than it already is. I personally would consider everyone as close to crazy if he should choose Pentium 4 plus i845 and PC133 SDRAM.

Writing

[guacamole]

Quite frankly, the guy who wrote the article should take a basic "reading and composition" course at a local community college because the quality of that article is not better than your average useless slashdot posting. I wanted to vomit while reading it.

Re:Writing

[jcarley]

He writes in a very idiosyncratic manner but it is hardly bad. These are personal reviews rather than a commercial endeavour of his. If you don't like them, don't read them.

Trivia

[Daniel Rutter]

Howdy. Page mine. Server looks fine to me. Site hosted in USA, not here in Australia. So if you can't see it, not my fault. And I can get my mail just fine, thanks :-).

On a more interesting note, I put that review up on the 30th of August, which was while motherboard manufacturers were still getting busted for even saying that they'd shipped i845 boards, because the chipset hadn't officially been launched yet.

But here in Australia, for some reason, the boards were already being sold retail. I just grabbed those two from m'verygoodfriends at Aus PC Market.

I should probably update the review; I bet Abit and Asus have product pages for those boards, now :-).

Good line

[Argy]

"The kind of people who manage to cram three syllables into the word 'Athlon' are, most likely, not going to buy one."

The author could use a grammar checker for subject/verb agreement, but he does have an amusing writing style, considering this was a motherboard review.

Re:Good line

[Daengbo]

So..............I teach (English) for a living, and everything looks fine. Subject = people / verb = are....????????

Re:Good line

[at_18]

Subject = "The kind" / verb should be "is"

Re:Good line

[Argy]

Sorry for the confusion, I didn't mean that line in particular needed a grammar checker. The article as a whole, however, had a number of errors.

Re:Nocturnal Slashdotting

[friday2k]

COuld it be that /. readers are all over the world. When you close your tired eyes somebody will be sitting at work, in school, at home and be pretty awake to read /.
Just a thought, you know ...

Re:Reviews are cool, but whats the best hardware n

[EulerX07]

For AMD: I'd say don't buy anything until the KT266A (the A being VERY important) chipset is used by the manufacturer. Personally I'd wait until Asus churns out an A7V266 that uses the KT266A. The performance difference is staggering, it blows out of the water the Sis 735, which in turn outperforms the KT266 and AMD760 on mostapplications. Nforce reference board should be available soon for the benchmarkers, so we'll see what the 2X memory bandwidth does for the athlon.

I'm not enough of a whore to go out and find the links to kt266a reviews though.

Ooh...neat!

[Jin Wicked]

I remember you! I wrote you probably about three years ago about some page you had with "warning signs" for troubled teenagers, with the army boots, and the (gaming?) magazine, etc etc. I wanted to know if that was really you in the photograph.

I think I found you from something to do with about 300 (1000?) sparklers being wired together to create a huge bomb sort of thing...

I should try to find that again...

Okay sorry...just a neat coincidence. I'll shut up now. :) (Please tell me I'm not going crazy!)

Re:Ooh...neat!

[Daniel Rutter]

> I remember you! I wrote you probably about three years ago about some page you had with "warning signs" for troubled teenagers

That'd be this page...

> I think I found you from something to do with about 300 (1000?) sparklers being wired together to create a huge bomb sort of thing...

That'd be this page (and this one)...

> (Please tell me I'm not going crazy!)

You aren't. Well, not any crazier than an artist with a top hat habit is already likely to be.

Where should I buy an Athlon

[JeffL]

This is serious, and hopefully not too offtopic. I purchase lots of desktop machines, and for myself I can build exactly what I want from whatever pieces I think are best. However, for some random new post-doc I just want a decent machine that is easy for me to buy, and easy to get fixed if it breaks.

I have always bought Dells, because they make doing educational institution purchases incredibly easy, and if I need service I just call one place. I can customize the computers I am buying, and their prices are reasonable.

I am finding myself in the position of having to buy a very fast computer for somebody else. The problem I am running into is that Dell does not sell Athlons. I can buy a 1.8Ghz P4 from them for about $1900 fully loaded. I can also build myself a dual Athlon 1.2Ghz for the same price, and the Athlon is much faster.

So my question, is there a reputable and reliable company which sells customizable Athlon machines for a reasonable price?

Re:Where should I buy an Athlon

[seregine]

Try here:

http://www.cyberpowersystem.com/

(I don't represent CyberPower, but I am planning to buy an Athlon PC there soon...)

Mikhail

Re:Where should I buy an Athlon

[ZeroConcept]

Any1 knows where can buy a dual TBird system?>

Re:Where should I buy an Athlon

[Animats]

HP, Compaq, and Sony all offer AMD CPUs if you want a prepackaged system.

Polywell Computers in South San Francisco is a good system builder for high-end systems, and they sell AMD CPUs.

Re:Where should I buy an Athlon

[vovin]

The first two offer custom builds with the Tyan S2460 (cheaper) board. I don't know if a2z will offer non-MP chips with the S2460, I don't know why not ... I bought my dual celeron (Abit) from there a while ago -- before the copper mine.

http://www.spartantech.com/
http://www.monarchcomputer.com/
http://www.a2zcomp.com/

I am currently getting a dual AMD 1.4 system from spartantech.com though the S2460 is on back-order :-(.

Monarch and Spartan are similar in over-all cost, I thing Spartan has a larger variety of components though.

Re:Where should I buy an Athlon

[DaveWood]

I recently price hunted quite a while for a friend who wanted a good workstation for doing 3D modeling. Dell was my first stop too. At the time I looked (several months ago) they did sell Athlons, and while their prices were at the reasonable end of the big 1st tier builders, I found it was possible to beat them by at least $400 with the 2nd tier vendors.

We built what, at the time, was the fastest uniprocessor x86 computer available (1.4Ghz Athlon, DDR) for ~$850 (not including monitor). The URL is:

http://www.epcworld.com/

I see that today, the same system (which has - barely and arguably - been edged out of the "fastest possible" title by the ludicrously priced 2Ghz P4) is now selling for $783.

We found them to be acceptable, though not thrilling to deal with, and the hardware was of good quality.

Glad I'm not a gamer or running Windows

[PD]

I've got Linux on my Celery 300 (overclocked to 333 - couldn't get it any higher without cooling and it's more hassle than it's worth). With 320 megs of ram and a lot of disk space, I can't imagine what I'd do with a bigger computer. I also have a Thinkpad laptop with an 800 MHZ Pent III and as far as system usability goes, I cannot tell the difference at all. X is fast, compiles are very fast on both systems.

So, I will just sit back and laugh while I use my trusty Celery 300 for the next 5 years or so. Maybe then I'll pick up a real cheap antique Athlon or something to replace it.

Re:Glad I'm not a gamer or running Windows

[pjgunst]

interesting... but some people are screaming for some faster number crunching CPU. As a (amateur) developer, most of my CPU cycles are wasted on compiling code or running sanity checks. That's quite unproductive, don't you think?
That's where a i845 comes in handy. Really, I don't mind people calling me an idiot, I just need a reliable, cheap and relatively fast machine and I couldn't care less about memory bandwith. Since I don't have that much cash to burn, I'm delighted these kinds of chipsets exist.
I can replace my desktop machine for considerably less and upgrade more often.

Unless there's a comparatively reliable and cheap solution on the market, I consider myself a stupid. Chipset for the stupid? Chipset for the cheap bastards who don't need a mean lean machine.

Re:Glad I'm not a gamer or running Windows

[Angst Badger]

I am running Windows on a machine not much faster than yours and with less RAM, and it works just fine. The most processor-intensive stuff I do is run Photoshop under Windows and gcc under Linux, and while I wouldn't mind being able to do a kernel compile in a few minutes like I can on the dual 1GHz babies I manage at work, then again, I don't recompile the kernel all that often, so who cares?

Years ago, I figured we'd eventually reach the point that machines would be good enough for the average user, and eventually good enough for me, too, and then upgrading would slow way down. It seems most of us have reached that point. Gamers, as you note, are the obvious exception, but even if I had the time to play games, the main obstacle there isn't processor speed, it's the expensive and not especially well-supported graphics and sound cards. (Obviously, someone who's really into gaming will find this less of an obstacle than I do.)

My next "upgrade" is less likely to be a new desktop machine than it is a household file server with four 80GB IDE drives so my wife and I can share MP3s across the household LAN. And for that, any cheap-ass second-hand machine will probably do just fine.

Re:Glad I'm not a gamer or running Windows

[PD]

1) I've tried overclocking. Random lockups at any speed over 75 MHz. Too bad. It's an IWill BX motherboard, and I've got the PC-100 RAM.

2) The laptop is a thinkpad with 320 megs of RAM. Once the compiler is loaded into the cache, my hard drive light barely flicks. It's also a top of the line Thinkpad, so it's NOT a slouch in that department anyway.

ummm

[BEA6D]

how come not all the images load???

Anandtech covered this much better 2 weeks ago

[Binary Tree]

http://www.anandtech.com/mb/showdoc.html?i=1533

Insightful? No... Dumb!

[Dr. Spork]

1. It's not a virtue pay Rambus money.
2. DDR makes most apps run faster than Rambus on the P4.
3. What CPU heavy applications do you use on a regular basis which have full SSE support?
4. Commitment to Open Source from Intel, not AMD? Don't make me laugh! Oops, too late... thinking of Wintel.

I guess there will always be people like you, eager to spread FUD which has repeatedly shown to be erroneous. P4 FUDsters inevitably retreat to "reliability" because that's the only thing that doesn't get measured in benchmarking tests. (On every significant thing that has been measured, you lose.) So why are you so damn sure about this? Or are you just repeating this because yo mama told you to buy Intel, or because you heard this in church? Maybe you're hoping for a miraculous share price recovery... I wouldn't...

Re:Insightful? No... Dumb!

[Detritus]

Grow up. Your dick isn't any bigger if you are some kewl d00d running Linux on a liquid helium cooled Athlon.

There are perfectly good reasons to buy an Athlon or a P4. Which one to select depends on what your needs are. There is no "wrong" choice.

Re:Insightful? No... Dumb!

[EnglishTim]

Idiot.
Did you even bother to read his post properly?

1) He never said it was a virtue. He just said it has higher bandwidth. He's right.
2) He mentions he's running quad processor configurations, which means that he's going to be very dependant on memory bandwidth - hence the Rambus memory.
3) '[his] company's heavy data processing algorithms'... Yes, people *do* write in-house code, you know...
4) Hoisted on your own petard on that one...

FUD? where exactly is the FUD in his post? As compared you yours? In his experience the Athlons are less reliable. It wouldn't suprise me - several people have mentioned reliability in this topic. Do you think he's making this up? Do you even *have* any experience with P4 systems?

Yes, Intel is normally taken as the 'big bad guy' of the processor industry. Yes, Rambus has undoubtably been very nasty with their patents, but all of this has nothing to do with whether or not P4s + Rambus memory are actually any good or not.

Next time, think before you type.

Re:Insightful? No... Dumb!

[PurpleFloyd]

Did you bother to read the post properly? If you would realize, there is no multiprocessor support for the P4, and probably will not be for quite a while (they have to redesign a lot of the chip's architecture). He is a fscking troll. You believed him, get over it.

Aargh. Chipset problems.

[Bollie]

I've got an Athlon 1.2 with an AMD761 northbridge and a VIA vt82c686B southbridge.

Guess what, it's got a bug on the southbridge that causes anything on the PCI bus to get corrupted when an SB Live is installed. Aargh!

VIA keeps blaming Creative and vice versa. Good grief people! What happened to standards!

It was my assumption that in order for a card to be PCI compliant it had to pass certain tests. Same thing with a bus controller.

The moral of the story is the following: AMD makes nice CPUs but the chipsets that support them suck. Oftentimes. Intel makes sucky CPUs but their chipsets are nice. Oftentimes.

Lets hope the sucky chipsets Intel are introducing causes the AMD support chipsets to magically improve. Hey, anything can happen!

Re:Aargh. Chipset problems.

[Warin]

I've always heard about this problem, but I find it strange that I have been running my MSI K7 Master (AMD/VIA Hybrid) with an SB Live for two weeks now and havent had a single problem with the PCI bus.

Stick to reviewing the motherboard, Dan

[Pov]

Unfortunately, this article strays from reviewing the motherboard to take shots at Intel. Now, normally I have no problem with this, but it fails to take into account the new instruction set the P4 is designed for while the AMD's xHammers offerings are still running on the old PIII-level and below instruction set. In tests I saw reported in PC Magazine the P4 destroyed AMD in floating point calculations (though it was only narrowly better in integer math).

Now of course this isn't going to make any difference on most current software, but if you're a business or individual looking to cash in on high processor speeds that won't be caught by slowly advancing software, then the P4 may be your way to go.

There are a number of good things about the P4's new instruction set and architecture like 128 integer and 128 floating point registers, not to mention making use of predication and data speculation at the hardware level.

This guy should have stuck to the motherboard instead of trying to attack Intel. They may actually be doing something competitive other than being huge and having vendor buddies this time around.

Re:Stick to reviewing the motherboard, Dan

[dutky]

There are a number of good things about the P4's new instruction set and architecture like 128 integer and 128 floating point registers, not to mention making use of predication and data speculation at the hardware level.

um, have there been some really big changes is the IA32 or is this guy getting P4 confused with the IA64?

Re:Stick to reviewing the motherboard, Dan

[at_18]

...the new instruction set the P4...

There are a number of good things about the P4's new instruction set and architecture like 128 integer and 128 floating point registers, not to mention making use of predication and data speculation at the hardware level.

You are talking about the Itanium (IA64)!

The P4 is a strictly x86 architecture, designed to reach crazy clockspeeds (hence the long pipeline).

The most interesting thing about the P4, speaking of chip architecture, is the trace cache. Basically, the L1 instruction cache is replaced by a micro-op cache, saving two or three pipeline stages and some silicon (the three x86 decoders in the Athlon are VERY complex and play a big role in achieving its performance. Something like a trace cache would benefit a lot, IMHO).

DOS?

[FordImperfect]

Coming to think of it... i can use slashdot as a perfect DOS tool... don't like a site? or its owner?? Submit it to slashdot.END.

Of course the simple solution...

[Arker]

...would be to put the cooler on right the first time. :)

Your site's up? I think not.

[Romancer]

I just tried it and I still get the no data error.
Might want took at the serverlogs.

Re:Yawn..

[tenman]

I thought it was somewhere in the 30% range? oh well

Site blocks NS4

[ChrisCampbell47]

His site is unreadable to visitors using NS 4.0x, and possibly to visitors using NS 4.x .

I get a "document contains no data" popup error.

Works great in Mozilla. Perhaps his webserver doesn't know what to do with my client ID string? There should be a fallback position.

Dan, please make your website complaint [enough] with standards so that all browsers can at least see the basic text. Thanks.

Re:Site blocks NS4

[ncc74656]

His site is unreadable to visitors using NS 4.0x...

Dan, please make your website complaint [enough] with standards so that all browsers can at least see the basic text.

Seeing as how Nutscrape has a problem implementing standards properly (its CSS implementation blows goats), I don't see how you could do more than a basic design without either (1) breaking all the rules to make a site that renders properly in Nutscrape or (2) make a site that follows established standards, and screw the people (both of them) who are still using Nutscrape 4.x.

A third way would be to detect the browser and send either a standards-compliant page or a "lobotomized-for-Nutscrape" page. I did this in the redesign of this commercial site and refined it a bit further when I redid my personal site. It's not that I personally care if people who continue using outdated, buggy software can access my site...for the dot-com site, accessibility was considered important enough to figure out a work-around.

Here's a test for you: pull up my site in Nutscrape 4.x and in another browser (Mozilla, IE, Lynx, Opera...it doesn't matter). Save the returned HTML (grab the stylesheet, too) to a file somewhere on your webserver and have W3C's validator check both. You'll see that one validates as HTML 4.01 Strict, while the other doesn't validate as anything. Now load the page that validated properly into Nutscrape and tell me what you get. It's a mess, isn't it? It displayed just fine in your other browser, though (unless your other browser was IE 2 or something similarly ancient).

Re:Site blocks NS4

[Daniel Rutter]

> His site is unreadable to visitors using NS 4.0x

Or, to put it another way, no it isn't. I just read a few pages in Navigator 4.08 for Windows, no problemo.

Mind you, I've had the occasional e-mail from people telling me that there's some magic cookie in my HTML that stops _Mozilla_ from rendering it properly.

I just can't please you people, can I :-)?

Nforce and Dual Channel Memory!

[mprinkey]

The use of a dual channel memory implementation is a very important issue that motherboard and chipset manufacturers have been glossing over. Nvidia is revisiting the concept of dual data channels for memory access with its Nforce chipset in addition to using DDR RAM. IMO, that is the direction that both Intel and AMD chipsets should take.

RDRAM has exceptional memory bandwidth, but it will be equalled or exceeded by the NForce dual channel DDR offering. Moreover, Rambus only provides this bandwidth by using dual channel implementations.

I do hope that Nvidia seizes the chipset market from half-assed players like VIA, or at least forces the rest to get their acts together. Nvidia needs to roll Athlon MP support into that chipset and set the whole market on it ear. Based on benchmarks I have done for high performance computing applications, the Athlon succeeds now *in spite of* the chipset and memory architecture. The P4/RDRAM is the better choice for many of these applications because of memory bandwidth limitations in the VIA/AMD DDR implementations. The same is doubly true for the Athlon MP motherboards, while the Intel 860/Xeon/RDRAM combination provides enough memory bandwidth to satisfy two P4s.

I am glad to see Nvidia setting the pace here. They are experts at getting *real* performance out of cutting-edge memory technologies. I expect the Nforce to deliver, unlike the lackluster DDR implementations we've seen from VIA and Ali.

Re:Reviews are cool, but whats the best hardware n

[spiro_killglance]

The Nforce Reference board was Benchmarked today
at Anand's,
runs about par to slightly slower than KT266A,
Maybe a few tweaks from Nvdia was boost it above
the KT266A.

Umm there's no numbers in that review

[esses]

It says the equilant of "It has SDRAM, so it must suck"

I can't personally forsee a P4/SDRAM chipset working circles around P4/Rambus or an Athlon system, but they don't suck.

Darn it

[EnglishTim]

I did read the post properly, however, I didn't realise that the P4 had no MP. Okay, my bad.

HOWEVER, the guy who replied to it made no decent points in his post either. I think my points still stand, in a kinda hypothetical way...

what we really want...

[Cheetahfeathers]

Why isn't something like this coming out in the US?

These are the types of systems I would _really_ like to have.

Low powered crusoe systems would rock, for everything I do at home.

Need two kinds of help.

[twitter]

Does truth count? I've consistenly pronounced that word, "Ath-a-lon", but I have bought one. Oh well, at least I know it's "new-clear" not "nuk-u-lar".

I also know that AMD K6/2s at 500MHz run Debian well, and a 650MHz Athlon clasic is pleasantly fast. Knowing that Tiger direct will dump an 800 MHz clasic and mobo for $90 has me sorely tempted to upgrade a 130MHz pentium toy box.

So there you have it. Someone who's pronunciation is just awful with too many boxes around, unable to restrain his spending. Thank you for fixing one small pronunciation problem.

PIII prices rising?!

[FatSean]

Not according to pricewatch. I hope they don't go up, I was hoping to buy a pair of 1GHz chips for chepa in a few months.

Pure FUD

[DABANSHEE]

I've found AMD chips to be as reliable as Intel chips.

The main reason why so many corporates go for Intel over AMD is that they've been conned by the FUD that eminated from the 486 cum 586 days.

The fact is virtually all the corporates who say these sort of things have never used an AMD based system since the 486 cum 586 days.

Plus how many corporate IT managers & decision makers even know what they are talking about?

Most of them have never even built a system themselves.

Gez, a mate of mine worked in the drawing dept in a huge international construction company & they were about to go P4 1.7 GB route, because they just took what the Intel people said as gospel.

So I sat down with the IT blokes there & explained the various pro 'n cons.

Such as the fact that unlike generational upgrades in the past, where clock for clock things improved (for example a p5 based core like the P70 is faster than a 100mhz 486, & a p6 based core like a Pentium Pro 200 is much faster than a P5 based Pentium 180, especially on pure 32bit code), the opposite was true for the p7 cored P4, clock for clock they are abysmal compared with the even the p6 based Celeron - they honestly thought that as it was a generational upgrade that clock for clock the P4 was faster.

Also I explained that AMD based sytems are just as reliable, pending decent mainboards.

So I made up a 800mhz Celeron system (Intel discourages corporates from using the Celeron) & a 1.2GB T'bird system, for them. So they could spend 6 weeks comparing them with the P4 demo system they were given.

They realised I was right, & in the end went with 8 new Pentium !!! based systems, that matched the exact specs of my Celeron demo system, except for the CPU (to cover what they needed straight away), & purchased two 1.2GB DDR-SDRAM Athlon systems for long term evaluation (there really is extreme prejudice against AMD in the corporate sector - it's funny the IT network 'hardware' admin staff will make AMD systems for themselves & their relatives, but the management & 'software' staff seem to have almost idealogical opinions against them).

They are extremely grateful now, as they will save many thousands when they do eventually make their decision - they definitly won't go with the P4, especially now Intel's about to change socket formats again. They are now debating between Celeron, P!!!, T'bird & the new AthlonMP CPUs.

I think they'l go the P!!! route, as many enterprise suppliers make it hard for businesses to chose Celeron or AMD systems. Even so the P!!! is much better than the P4 (unless one is using optimised code, but then its only now we are even starting to see code that's P5+ optimised), & heaps cheaper too.