Athlon 64 3400+ Reviewed

SpinnerBait writes "Unlike the Athlon 64 FX-51, this new 3400+ rated Processor, has a 64 bit memory interface, with its integrated memory controller, drops in at several hundred dollars less than an FX-51 and is also clocked at 2.2GHz. It gives a P4 3.2GHz Canterwood based machine a run for its money too, as this review with benchmarks at HotHardware reports. And where is Prescott? Fortunately for AMD, it's a bit tardy to market and this will give this new Athlon 64 speed bin time to take a firm hold."

Comments

It figures

[fataugie]

I just bought a Barton 2500, now I bet the prices are going down as I type this.

Good news, at least now I have an upgrade path for a little longer....

Re:It figures

[fataugie]

Nevermind....sorry, I misread it as an XP3400 (32 bit, not 64)

Re:It figures

[Anonymous Coward]

You're insane. 2500 bartons have been $90 for a LONG time. You should have got one ages ago.(well... half a year, anyway)

I'm ready give mine to the kids and swap it for an A64 3000.

Nice how these all coincide

[Anonymous Coward]

Glad manufacturers get their butts together! Athlon 3400+ compared with the 2.6GHz G5 just announced at macworld, I wanna see the mac zealots vs the x86 weenies in a table match!

one two three go!

Re:Nice how these all coincide

[Anonymous Coward]

Announced when? At the keynote? I thought that hadn't happened yet...

Re:Nice how these all coincide

[godzilla808]

Hasn't.

Re:Nice how these all coincide

[Anonymous Coward]

The problem with the G5 is that the upgrade path is much more costly. You just can't go to Pricewatch and drop in a new motherboard.

what are speed bins?

[Anonymous Coward]

"this will give this new Athlon 64 speed bin time to take a firm hold"

What's a speed bin?

Re:what are speed bins?

[Anonymous Coward]

It's a bin for holding your speed. AMD's new processors let you save up cpu cycles for a rainy day.

Re:what are speed bins?

[Anonymous Coward]

Something like this?

Re:what are speed bins?

[Anonymous Coward]

Why is this moderated as 'funny' and not 'ignorant' :-P ?

Processor makers 'bin' processors. That is, they try for the fastest speed, but if the chip doesn't make it, it get's 'binned' dowm the line and tried as a lower-speed chip. They can also 'bin' due to market-reasons (putting hi-grade chips in the low-speed bin because of demand, etc)

Re:what are speed bins?

[Tony Hoyle]

It's a bin that fills up *really fast*!

Re:what are speed bins?

[kfg]

What's a speed bin?

It's almost, but not entirely, unlike a Wall of Death.

KFG

Re:what are speed bins?

[hab136]

"this will give this new Athlon 64 speed bin time to take a firm hold"
What's a speed bin?

In case you're not trolling, chip manufacturers crank out one design of chip, test it, then put them into bins based on how fast they can run reliably. They probably don't actually use plastic bins, but you get the idea.

Thus, a "speed bin" - a lot of chips designated to run at a certain speed, despite the fact that it's the same design and metal as a chip designated to run at a slower speed.

Re:what are speed bins?

[Anonymous Coward]

Due to minor process variations (random faults in materials, equipment, background radiation etc.) every manufactured chip is different. Some chips work fine at higher speeds, some chips only work properly at lower speeds, some chips fail to work at all. Since these microprocessors are at the cutting edge of silicon process technology, the variation matters. Now, to sort out which chip works at which clock speed, the manufacturer has to test every chip and classify them accordingly. Some are sold as 2 GHz chips, some are sold as 1.8 GHz chips, and so on. These different grades are called "speed bins".

Re:what are speed bins?

[CubicDDD]

moderating "ignorant"?
Maybe your ignorance could be moderated according.

Could be it was an honest question. Not everyone in the world uses english as their primary language. (including myself)

And even then there could be problems. (see dunno-what-speakin'-Bush: "mis-underestimating")

Re:what are speed bins?

[Anonymous Coward]

Why is this moderated as 'funny' and not 'ignorant' :-P ?

I'd hardly call not knowing what a speed bin is "ignorant". The poster didn't know, wasn't trying to be funny about it (AFAICT) and isn't responsible for the moderation it got.

Re:what are speed bins?

[LiquidCoooled]

OK, you lot seem to be way over analysing this.
it looks like a simple bloody typo:

"this will give this new Athlon 64 speed bin time to take a firm hold" ..becomes..

"this will give this new Athlon 64 speed in time to take a firm hold"

i doubt it has anything to do with grading the processors or anything......

Re:what are speed bins?

[Rhubarb Crumble]

I'd hardly call not knowing what a speed bin is "ignorant". The poster didn't know.

That's what "ignorant" means. From latin, ignorare, "to be unaware [of sth], not to know". Antonym of scire ,"to know, to be aware [of sth]". (cf. "Science").

Re:what are speed bins?

[Anonymous Coward]

Yes and no. Ignorancy means that, but moderation of "Ignorant" generally has additional negative connotations. This means that such moderation might seem misplaced for the original question.

Furthermore; while poster was ignorant of the fact, the posted question itself was not ignorant per se. Moderation is applied to comments, questions, not to posters.

Re:what are speed bins?

[Rhubarb Crumble]

Yes and no. Ignorancy means that, but moderation of "Ignorant" generally has additional negative connotations. This means that such moderation might seem misplaced for the original question.

There is no moderation "Ignorant" (at least there wasn't last time I had mod points)...although many feel there should be.

Agree about the connotations (i.e. "you SHOULD know this"), but that's another matter.

Furthermore; while poster was ignorant of the fact, the posted question itself was not ignorant per se.

Well, a question is not capable of knowledge, and therefore can neither know nor not know. But the question was posed by a person who was ignorant of a particular fact and concerned the matter he/she was ignorant about.

I don't know if there is such a term as an "ignorant question" (which would be the opposite of a rhetorical question), but if there is, it qualifies.

Not that I'm saying the poster is stupid for asking - it's a fair question.

Re:what are speed bins?

[Anonymous Coward]

or alternately, you could read "bin time" as a phrase, you know, the cpu needs to spend time in the bin at the store before people buy it. must say I like the typo idea more. "speed bin" makes no sense in this context.

Re:what are speed bins?

[Slime-dogg]

Perhaps, if you knew what ignorance meant, you wouldn't use it in an incorrect way. The parent post of your comment knows what ignorance means, but it seems that you do not.

Being ignorant isn't a bad thing. There are a lot of things that I am ignorant about. I would probably prefer to be ignorant of them, if I ever found them out.

The word "ignorant" has been misused, typically within the race card. It probably started as "you are ignorant of what my people went through," but it somehow morphed into a cuss word akin to the "N"-word.

Re:what are speed bins?

[Slime-dogg]

I thought "speed bin" was just the long name of "/sbin." Yeah... the kind that every user should share.

Re:what are speed bins?

[antiMStroll]

Then everyone is ignorant since none of us know everything. Any term applicable to all loses it's differentiating power and becomes useless, so I'm guessing the grandparent post meant "ignorant" in a more limited, less pleasant way.

Re:what are speed bins?

[arodland]

Perhaps you should think on the meaning of "ignorant".

If it was an honest question, that is PRECISELY what ignorance would be.

See also the age-old adage "ignorance is excusable. Stupidity is not."

Re:what are speed bins?

[tunah]

In case you're not trolling, chip manufacturers crank out one design of chip, test it, then put them into bins based on how fast they can run reliably.

Trolling? Why does every post have to fall into the dichotomy of trolls/karma whores?

Re:what are speed bins?

[ccp]

Trolling? Why does every post have to fall into the dichotomy of trolls/karma whores?

+1, Insightful.

I'm bothered by Moderator points all the time, except when I need them.

Cheers,

Java VM is what we need

[thammoud]

The Itanium is too expensive and slow. Ditto Sparc. AMD 64 bit servers running 64bit Java VMs will make for a killer combination.

Re:Java VM is what we need

[maj_id10t]

Shouldn't bee too long with Sun's announced support of AMD chips. They already have it for the Itanic. http://java.sun.com/j2se/1.4.2/install-linux-64.ht ml

Re:Java VM is what we need

[Anonymous Coward]

A 64bit OS that 90% of the world can and will use would, of course, be more important.

The people who need 64 bit are in the following order:

High traffic servers needing more than 4 GB of RAM

Designers using 3D modelling (CAD, 3dsm, etc)

Computational clusters

End users

You see, the first bullet can use Solaris, FreeBSD, Linux, or OS X.

The second bullet ... well, they're still waiting on Windows.

The third bullet is FreeBSD and Linux.

The fourth bullet ... still waiting on windows.

Clearly, more important than the jvm is a version of Win2k3 that supports the 64 bit memory extensions.

Re:Java VM is what we need

[Bert64]

Alpha (despite not being developed for years) and Power are still very much competitive.

Re:Java VM is what we need

[Juergen Kreileder]

Blackdown has released J2SE 1.4.2 for AMD64 a few weeks ago, see http://www.blackdown.org/java-linux/java2-status/j dk1.4-status.html.

Re:Java VM is what we need

[ProtonMotiveForce]

Oooh, sorry. In the real world, as opposed to your make believe land of fairies, Itanium is anything but slow. Was the last model you looked at the original 800Mhz model?

Newer itaniums _smoke_ anything the Opteron can do in their applicable domains (engineering, scientific, database, etc...).

Now expensive I won't argue with.

Re:Java VM is what we need

[Anonymous Coward]

yeah, a RAM killer.

Re:Java VM is what we need

[Anonymous Coward]

umm, typical intel zealot, SMTA.

Re:Java VM is what we need

[Master Bait]

I wouldn't call it _smoking_ but the truth about the Opteron is commoditization of the 64-bit space. Wasn't that supposed to be the impetus behind Itanium in the first place?

Re:Java VM is what we need

[Anonymous Coward]

So, video compositors don't need 64-bits?

Re:Java VM is what we need

[ProtonMotiveForce]

*cough*Stupid*cough

Typical shit-for-brains monkey.

Re:Java VM is what we need

[p3d0]

Hey, I'm doing my best, but that's about all I can say. :-)

(I'm the lead developer on IBM's AMD64 JIT compiler porting effort.)

Re:Java VM is what we need

[tunah]

The Itanium is too expensive and slow. Ditto Sparc. AMD 64 bit servers running 64bit Java VMs will make for a killer combination.

Or maybe they'll be slow too. I think you're missing something...

Just kidding, I love java :-P

Re:Java VM is what we need

[stock]

" AMD 64 bit servers running 64bit Java VMs will make for a killer combination."

Please piss off with your Java. We have linux with c and c++ code, which runs on all major platforms. No need for Java here. Linux and GNU software itself is the best "Java" ever written.

Robert

anadtech is my fav site for reviews

[McVeigh]

http://anandtech.com/cpu/showdoc.html?i=1941

Re:anadtech is my fav site for reviews

[ehb]

Why not go straight to the one-page printable version?

http://anandtech.com/printarticle.html?i=1941

Re:anadtech is my fav site for reviews

[Anonymous Coward]

funny enough, i cant view the benchmark results in the mentioned review because im on a dual opteron 64bit linux system where there is no flash plugin for mozilla firebird available.

ah .. the cost of properitary web standards that emerged on windows ...

REALLY cool chip?

[GeckoFood]

Found this little gem in the article:

It kept our CPU running in the mid -40C range while gaming at default clock speeds.

Last AMD I had ran hot enough to roast a turkey from 10 feet away. -40C would freeze it solid.

Re:REALLY cool chip?

[DreadSpoon]

I believe the "-" was meant as a separator, like how one would say mid-range.

Re:REALLY cool chip?

[Anonymous Coward]

Found this little gem in the article: It kept our CPU running in the mid -40C range while gaming at default clock speeds.

P4's still run a lot cooler

Re:REALLY cool chip?

[Anonymous Coward]

erm... 40C isn't exactly in the freezing range, unlike 40F...

Re:REALLY cool chip?

[Anonymous Coward]

... but -40 C is exactly the same temperature as -40F !

Re:REALLY cool chip?

[Anonymous Coward]

There was the one super cooling article with an Athlon 3200 cooled cown to -80C. Used a two tiered refridgeration system I think.

Re:REALLY cool chip?

[jpr1nd]

erm, unless your processor is melting out of the socket and running down the motherboard it is frozen :\

it is solid isn't it?

Re:REALLY cool chip?

[arty3]

Ehh.....Last I checked, turkeys were already solid objects.

Misunderstanding the metric system?

[Inoshiro]

"We found the heatsink to work quite well. It kept our CPU running in the mid -40C range while gaming at default clock speeds."

If your CPU runs at -40C, you have something very special. I, for one, would be worried about condensation from water becoming ice on contact with the CPU at that temperature!

Re:Misunderstanding the metric system?

[GeckoFood]

No misunderstanding the metric system here. -40F happens to be the same as -40C.

Re:Misunderstanding the metric system?

[Pingular]

If your CPU runs at -40C, you have something very special. I, for one, would be worried about condensation from water becoming ice on contact with the CPU at that temperature!
It happens.

Don't forget Anandtech's review...

[cK-Gunslinger]

Anandtech

Looks like a winner to me!

Re:extra links

[eyegor]

Mod parent down. The tarrato link will redirect you to goatse.cx

Nforce 3 :/

[Mdalek]

The review in the article is slightly hampered by their choice in motherboard.

Compiler optimtizations???

[Shisha]

Well not that I'm buying one anytime soon, but it's nice to know that once I buy one, I'll get a Linux distro, that is compiled & optimized for a 64bit CPU. So for me only Mathematica will run in the 32bit (slower) mode. But Gimp, mplayer, video editing apps, hell even twm and xclock, will be compiled for 64bit CPUs.

I was wondering how is this going to be sorted out by application vendors on PCs? Are they going to release 64bit and 32bit versions? Is every CD going to contain both? What about 3rd party plugins? I've been asking the same question actually about Apple's G5, but www.apple.com (and I didn't search too carefuly) is bit short on nasty details like this. Is it really worth getting a 64bit machine without planning to use Linux?

Re:Compiler optimtizations???

[avenj]

Is it worth getting _any_ machine without planning to use Linux?

Re:Compiler optimtizations???

[Anonymous Coward]

Microsoft has Windows XP for 64-bit AMD chips in beta, and should have a release out soon. From what I saw earlier, they had a pretty good transition plan in place, with 32-bit programs being partioned from 64-bit ones. I imagine in the short term, companies will simply include both versions if they have a 64-bit version available. The chip is still going to run 32-bit programs about as fast as a 32-bit cpu at the speed according to benchmarks I've seen.

Re:Compiler optimtizations???

[cK-Gunslinger]

Is it really worth getting a 64bit machine without planning to use Linux?

Well, if you want the current top of the line 32-bit performance, why not? That's a bit like asking "Should a get this Super Duty Dodge Ram with the best towing capacity available today, but also includes an extra cup holder I might never use." If it has what you need for a reasonable price, why question the extras you might never use. It's not like the 64-bit-ness is truly "wasted" just because you might not use it. Those extra registers can't be used to feed hungry children in Elbonia or anything. They'll just sit there unused.

Re:Compiler optimtizations???

[discstickers]

Actually, Apple's in a good position here. Mach allows "fat" binaries (ie, more than one binary in a single application icon). So both version can be distributed together.

They did a similar thing around the transition to PowerPC.

Re:Compiler optimtizations???

[Espen]

Apple proposes: "Packaging your Optimizations:

Code that has been optimized for the G5 by simple re-compilation will run without penalty on a G4. If you have done more in-depth, G5-specific tuning (levels 1, 2 and 3) then you will in all likelihood want to provide a separate binary. In extreme cases, you may decide that you need only offer one version of your software that runs on Power Mac G5 computers only. However, you'll probably want to support most or all of the Macintosh product line, which means that you need to decide how best to deliver the right code to each of your customers. There are several ways to achieve this; the first is:

Create different versions of your software for each processor that you support. This requires that you maintain three parallel code bases, something you may not want to do.

It is possible for your software to query the computer on which it is running to see which processor-related features are available. You can design your software to isolate processor-dependent code and call the appropriate version as needed. This leads to two additional strategies for packaging your application:

For every function that calls processor-dependent binary code, have your code call the appropriate version. If such functions are needed frequently, using this approach may decrease execution speed and make your source code (cluttered with if...then constructs) less readable.

Isolate processor-specific functions into frameworks or shared libraries, then have your software load the appropriate version when it starts up. This enables you to write your main code without wrapping function calls in if...then constructs."

(from G5 Optimization)

Re:Compiler optimtizations???

[DreadSpoon]

Likely, vendors will only ship one version of their app, the 32-bit version, since AMD64 CPUs can run 32-bit and 64-bit. If the app doesn't need or greatly benefit from 64-bit, why bother with it? (The code will likely need porting to 64-bit anyway, especially by average Windows programmers who've never had to worry about different CPU architectures before.)

Apps that greatly benefit from 64-bit support may either be 64-bit only, or provide both versions. I recall reading that the UT developers plan on requiring 64-bit mode for their next-gen game content development tools, for example.

Generally, it'll have to do with the market. If an app can get away with a 32-bit version, it will have one, since most people will still have 32-bit CPUs in the forseeable future. Apps that can really get a boost from 64-bit will have the extra money put into developing a 64-bit version in hopes that it will sell more copies that way.

Re:Compiler optimtizations???

[Shisha]

"Well, if you want the current top of the line 32-bit performance, why not?"

I haven't checked, but won't a P4 system give me better "speed per dollar"?

For me personally I couldn't care less about speed. With me the "weakest link" it's usually my brain or ocassionally the internet connection.

What I would care about more is a silent and small (think book sized) system. When I say _silent_ (not just almost silent), I mean that it won't need a CPU fan, no power source fan and that it would be based around a 1GB compact flash card. I would quite like a decent (not great) graphics card. And a 1 gig ethernet port. That way the fileserver can be under the stairs.

P.S. "speed per dollar" is a not very scientific term, I agree. But I don't now how to amalgamate GFLOPS, data troughput, L1, L2 caches etc. into one thing?

Re:Compiler optimtizations???

[Anonymous Coward]

for me only Mathematica will run in the 32bit (slower) mode.

*sigh*

You fanboys just dont get it. 32 bit vs 64 bit has nothing to do with speed. The 32 bit mode that Mathematica runs in is not the "slower" mode. All of the performance increases in Athlon64 are due to architectural enhancements that are completely independant of the size of the registers...

Re:Compiler optimtizations???

[mattdm]

sure, but isn't having *more* registers helpful?

Re:Compiler optimtizations???

[MenTaLguY]

Well, the nice thing about AMD's 64-bit thing is that, unlike ia64, it's even quite capable as a 32-bit processor.

So maybe think of it as getting a very good 32-bit processor, with room to grow as more 64-bit binaries become available?

Re:Compiler optimtizations???

[balloonpup]

Sounds like you want a Via-based system, if you're not interested in speed. They still sell some x86 procs that don't require anything more than a heat sink...though they still need a power supply fan...

Re:Compiler optimtizations???

[cK-Gunslinger]

Well, the Anandtech review has several charts showing price/performance ratios for different scenarios. In everyone one of them, the Athlon 64 3400+ and P4 2.8C take first and second places. So I would guess that, depending on your budget, either of these will give you the best "speed per dollar." Although, I would like to see some of the Althon XP's compared for reference. I would be willing to bet that the Athlon XP 2600+ 333MHz/512K Barton CPU for ~$95 would rank pretty high on that chart, simply because of its amazingly low price.

Re:Compiler optimtizations???

[FattMattP]

I was wondering how is this going to be sorted out by application vendors on PCs?

They'll recompile and then sell it to you as an upgrade, as always.

Re:Compiler optimtizations???

[Anonymous Coward]

Download the source code and compile with march set to x86-64

Re:Compiler optimtizations???

[Zathrus]

When I say _silent_ (not just almost silent), I mean that it won't need a CPU fan, no power source fan and that it would be based around a 1GB compact flash card. I would quite like a decent (not great) graphics card. And a 1 gig ethernet port.

You can get by without a CPU fan (see Via as others state). Good luck on the no PSU fan. And basing it on a CF card? With a 1 Gb ethernet port? Why? The card can't possibly keep up to the port (particularly for writes, which is another issue -- if you put any kind of swap space on the card it'll be destroyed in no time).

If you want a good site for quiet PC stuff, check out Silent PC Review. There's a wide variety of posters there, from aiming at truely silent, low-end computing to quiet, high-end systems.

As for speed/money, read AnandTech's review. It's either the Athlon64 3000+ or P4 2.8C, depending on task.

Re:Compiler optimtizations???

[Abcd1234]

Actually, a simple recompile could provide additional performance, as the AMD-64 has a host of additional registers over and above the standard x86 set.

Re:Compiler optimtizations???

[be-fan]

Actually, no. Most the the improvements are due to architectural improvements only available in 64-bit mode. They have little to do with the fact that integer registers are now 64-bit, but you don't get them in 32-bit mode anyhow. 64-bit mode on AMD64 should be about 20% faster than 32-bt mode, Mathematica is running in "the slower mode."

Re:Compiler optimtizations???

[DrSkwid]

It
Sure
Is

Re:Compiler optimtizations???

[DrSkwid]

not quite

my 500Mhz EPIA server has no fans whatsoever.

the brick PSU is rated as 240v AC / 1.8A (12v DC 4.5A)

Re:Compiler optimtizations???

[Anonymous Coward]

The only application I know will be put out in 64-bit is Unreal Tournament 200x (capable of running on 64-bit Windows XP, due out ?). The cd will only contain the 32 bit version, but the 64-bit addendum will be available online. AMD says the re-compiled code for a 64 bit app will be larger than the 32-bit code, so I would guess we're looking at a long download.

Re:Compiler optimtizations???

[Anonymous Coward]

Just a nitpick but it's got nothing to do with Mach. The workspace selects the appropriate binary.

Re:Compiler optimtizations???

[the_2nd_coming]

you don't need a 64 bit app if the app does not need the memory. the G5 and x86-64 run 32 bit instructions just fine and even faster than 32 bit chips.

Re:Compiler optimtizations???

[the_2nd_coming]

the G5 runs 32 and 64 at the same speed.

Re:Compiler optimtizations???

[Coryoth]

You fanboys just dont get it. 32 bit vs 64 bit has nothing to do with speed. The 32 bit mode that Mathematica runs in is not the "slower" mode. All of the performance increases in Athlon64 are due to architectural enhancements that are completely independant of the size of the registers...

disclaimer: I am no expert on CPU stuff, I just do a lot of math computations.

If he's actually doing any serious work in Mathematica then 64 bit does start to matter. High powered math is one of the areas where having 64 bit word sizes can make a significant difference. Of course, that's not making much difference in terms of speed, but rather precision (unless of course you're having to do lots of nasty juggling in 32bit mode to handle the required precision).

Jedidiah

Re:Compiler optimtizations???

[be-fan]

Yes. And that is relevent to a discussion about the Athlon64 how?

Re:Compiler optimtizations???

[phoenix_rizzen]

You are talking about price / performance -- or getting the most bang for the buck. Since the release of the original AMD 386 clone chip, AMD has had the best price / performance ration. IOW, the AMD chips will always give you the most speed per dollar spent on it.

As for wanting a super-cool, super-quiet CPU, that has decent performance, you'll want to check into the Transmeta TM5800 and TM5900 series chips, or the VIA C3 chips. Both of these can run without fans, and are clocked up around the 1GHz mark. These CPUs won't give you the same performance as a P3/P4/Athlon at the same megahertz, but they aren't horrible either.

Re:Compiler optimtizations???

[Anonymous Coward]

Good for it.

Re:Compiler optimtizations???

[Anonymous Coward]

it is relevant to a discussion on 32 an 64 bit instuction speeds, which this thread is about you dope

Re:Compiler optimtizations???

[stanmann]

Why not just go with a network boot and a massive ram disk of some sort ... esp since with gigabit and a fast fileserver you should be gtg. 4 giga ram 3 dedicated to "local" 512 to some sort of smart network cache and keep userspace remotely... then all you have locally is system and open files...

Re:Compiler optimtizations???

[balloonpup]

Cool, good to know :)

Re:Compiler optimtizations???

[DreadSpoon]

Unfortunately, it doesn't work that way. You can't compile and get just the extra registers. You have to take it all, which includes changes to things size changes of certain data types. The software will likely need modifications for this if the developers never intended to run on non-32-bit systems.

Re:Compiler optimtizations???

[Schugy]

Just get a Athlon64 3000+ and run it at 800Mhz when idle. This means a consumption of ~35W and a powerful heatpipe + fan @ 0 rpm. You only get noise when you really use its power. This solution scales very well. A Via Chip hasnt any power even if you demand it.

Re:Compiler optimtizations???

[p3d0]

Simple: applications that benefit from being compiled for AMD64 will be. Applications that don't, won't. That's the beauty of backward compatibility.

Re:Compiler optimtizations???

[Abcd1234]

Oh yes, that's absolutely true. However, the original contention was that most apps won't see any performance benefits from being converted to x86-64 apps, and this isn't true, simply due to the additional registers provided.

Re:Compiler optimtizations???

[DreadSpoon]

Aaaah. OK, now I see, sorry. I get off on the wrong tanget *waay* too often. ~,^

Is there a desktop market?

[ViolentGreen]

Looks pretty good. I still don't think there is a huge demand to have these in desktops as of yet. P4s are still very powerful and still compete with AMDs 64 bit chips. Even the Athlons are enough for most people to play the newest games and all.

I don't think that most people do the really computer intensive tasks that would benefit from 64bit chips plus the lack of truely 64 bit software that will give them this advantage is a hinderance as well.

I think it will be 2005 or maybe even 2006 before 64 bit chips become the standard.

Re:Is there a desktop market?

[Feathers McGraw]

Of course, the whole selling point of the x64 technology is that it acts as a high-performing 32-bit platform even if 64-bit on the desktop isn't really necessary. The last time I checked prices (admittedly, that was a while ago), there wasn't a huge price difference between the Athlon 64 3000 and the Athlon XP 3200. The only big difference was the slightly more expensive Athlon 64 motherboard cost, but $30-$50 isn't too much for what I would consider a decent bit of futureproofing.

Re:Is there a desktop market?

[0123456]

"I don't think that most people do the really computer intensive tasks that would benefit from 64bit chips"

Everyone would benefit from switching because of the extra registers in 64-bit mode and the low-latency memory controller. Some people have said they got a 10-20% speedup just from recompiling in 64-bit mode without making any changes to their code.

Of course if all you do is run Word all day that will make little difference... but if all you do is run Word all day you'd probably be happy with a Pentium-II.

Re:Is there a desktop market?

[i_r_sensitive]

Yeah but, while all of that may be true, I still want one NOW!

Re:Is there a desktop market?

[Anonymous Coward]

Memory is the reason we buy 64 bit machines. I agree with you that 64 bit chips will become standard in the next couple of years.

Re:Is there a desktop market?

[ViolentGreen]

Everyone would benefit from switching because of the extra registers in 64-bit mode and the low-latency memory controller. Some people have said they got a 10-20% speedup just from recompiling in 64-bit mode without making any changes to their code.

Seriously though, how many "regular" computer users have access to the source code for their applications or would even know what to do with it if they did.

This is great for universities and research facilities that use either their own software or open source softeare. I'm sure a lot of people will buy it out of ignorance just because of the salesman or just to have the biggest and best. But before it becomes really mainstream, you are going to have to have the 64-bit windows (not sure if it is available or not) and a decent selection of 64-bit commercial software.

Re:Is there a desktop market?

[0123456]

"Seriously though, how many "regular" computer users have access to the source code for their applications or would even know what to do with it if they did."

Uh, that's irrelevant, as people will buy 64-bit versions of software for their 64-bit PC... and that software will run faster just because it's been recompiled with a 64-bit compiler that doesn't waste half the time copying data between registers and memory.

"But before it becomes really mainstream, you are going to have to have the 64-bit windows (not sure if it is available or not) and a decent selection of 64-bit commercial software."

Yes, but I thought that was obvious. I'm sure Microsoft are just gagging to sell us all yet more copies of Windows and Office just so we have 64-bit versions.

Re:Is there a desktop market?

[Atzanteol]

Seriously though, how many "regular" computer users have access to the source code for their applications or would even know what to do with it if they did.

There is a vast horde of us Gentoo users who are laughing like hell at you right now.

This is the nicest thing about distributing applications as source. Who cares what architecture I run on, if it'll compile it'll be optimized and work on my local hardware! Itanium, x86-64, x86, any will do. I'll just 'emerge -ev world' and everything will recompile and run just fine. :-)

Re:Is there a desktop market?

[nelsonal]

The reason AMD build a 64 bit chip. You have to understand the changes going on on the server markets. In short x86 caught up with proprietary architectures. Realize that it probably takes about $2 billion to develop a processor architecture (I'm probably low here, but you get the idea). Let's assume that 15 million server processors are sold annually. Sun ships about 300k servers a quarter and has 1/3 of the market, most of them are at the lower end (ie lots more 1-4 processor systems than 72 processor systems). If a design lasts 2 years, each of the three big architecutres have about 10 million processors to spread design costs over, or about $200 in R&D costs per processor.
Because of the PC market, Intel and AMD combine to sell 130 million processors annually (100 million for Intel and 30 for AMD. If they spend the same $2 billion on a processor design and their desings also last 2 years AMD can spread development costs over 60 million processors for about $30 per chip. If they can sell these chips into the server market, they have a $170 R&D advantage over the propretary architectures, even if consumers never use 64 bit features (buying them only because they are fast or cheap. They are quite willing to add something that allows them to compete quite effectivly in the server market even if it makes the chips cost a little more in the consumer market. This is why Xeon is selling like hotcakes, while Itanium lingers on the shelves.

Re:Is there a desktop market?

[ViolentGreen]

There is a vast horde of us Gentoo users who are laughing like hell at you right now.

I use gentoo myself. However, the vast majority of computer users in the world use windows. That's what I was refering to by saying "regular" computer users.

Re:Is there a desktop market?

[Anonymous Coward]

The only big difference was the slightly more expensive Athlon 64 motherboard cost, but $30-$50 isn't too much for what I would consider a decent bit of futureproofing.

It's not even that good at futureproofing considering the 754 pin package is dead on arrival. The Opteron motherboards are where it's at. Unfortunately I can't afford $700 for an FX-51 chip to play games.

Re:Is there a desktop market?

[Anonymous Coward]

and damn, here we are, like suckers, stuck in 2004. dammit.

Re:Is there a desktop market?

[Lumpy]

I still don't think there is a huge demand to have these in desktops as of yet.

for the casuah home ding-dong user? you are right.

for business and companiesthat depend on processing power? Ha! the amd 64's rock massively.

I replaced a dual Xeon box here at work that was the CGI station running Blender, povray and yafray and producing better graphics than a maya station next to it, and faster... with a dual Opteron using a 64bit compiled Gentoo install on it...

Then we recompiled the apps for 64 bit.

I am getting a 70% increase in rendering speed. I'm betting that with some optimization this could work even better... the Blender guys are working on that right now BTW...

a 64bit linux version of Maya? the company said "maybe 4Q 2004 for beta testing"

which is a shining example of why open source is the way to go.

businesses using the number crunching and processing power and are smart enoughto have embraced linux for the needs it can fill are all over AMD64 right now.

Re:Is there a desktop market?

[Xua]

As much as I like Gentoo it's not quite true, applications have to be portable too. For example you cannot access memory on address unaligned on its size on SPARCs (I fixed a mistake like this in an early version of mplayer). I am not sure about Itanium (it is likely it has the same limitation as SPARC), but ia64 and amd64 are still considered experimental in gentoo. And there are always issues with exotic hardware that exists on some architectures.

Re:Is there a desktop market?

[ducomputergeek]

I used to work at an architecture/graphics firm that's been running 64-bit chips since at least 1998. It was called DEC/Compaq/HP ALPHA chips running at 500mhz when the PII's were brand new at 350Mhz.

In fact they had 27 Quad 500Mhz Alpha chips with 4GB of RAM each configured into a renderfarm and one on one, still beat the replacement dual 2Ghz Xeon boxes which replaced them last year.

However, they condensed a 1500 sqr foot room full of servers and wires into two, and now getting a third, IBM blade unit in two cabinats for the price of just one of those Alpha servers a few years ago.

As for my current company, we're upgrading the video production unit to all G5's and they are great. We saw some dramatic improvement with the optimized FCP running on the dual 2Ghz G5 vs our older dual 1.25Ghz G4's.

We've built our company around Mac's from the ground up even with the horrible over priced XRaid for storage, but the amount of time we saved in having everything work together nicely has been worth the cost in dollars to us.

Re:Is there a desktop market?

[Anonymous Coward]

Dude - if you're using Blender and povray at work, I feel sorry for you. Tell your management to get off their cheap asses and buy you some real rendering software.

Re:Is there a desktop market?

[Lumpy]

I'll bite your troll.

hey dude... tell me how to legitimately get a nsa maya station in the door for less than $10,000.00

oh you CANT.

I am able to install new hardware, and install the 3d tools on every graphic artists desktop for 1/2 the price of one Maya professional license.

cince then, the CG graphics quality and quantity has delighted our clients and customers.

it's making us more money, saving us money, and able to use the newest technology to be way ahead of our competition...

so I fell sorry for you, stuck in a dying company with a dying attitude.

a customer happy with the CG from Blender and yafray is a paying customer. the maya station sits there unused lately cince it's inconvienent and the artist figured out hoe ti do things better in the new free tools.

we keep the maya setup around just so we can advertise that word on our spec sheets. :-)

9 more reviews here

[whovian]

http://amdmb.com/#News-7458 or linkified.

Reviewed by amdmb, HotHardware, Neoseeker, CPU Performance, Tech Report, Hardcoreware, Hardocp, Hexus, X-Bit Labs.

Just as good (?)

[syberanarchy]

is also clocked at 2.2GHz

Maybe I'm missing something (64 as opposed to 32 bit) but I'm running an OC'd 3000 to 3200 that clocks at 2.21 ghz. I got it for a good 175. Why would I want to upgrade yet?

Re:Just as good (?)

[Anonymous Coward]

Because a good majority of IT professionals don't have time to deal with instability, testing, high-performance cooling, or nonsense like that. In the Real World, I need a machine that can render reliably on a daily basis. Overclocking is fun, at home, as a hobby. In the office from 9 to 5, machines need to come out of the box and "just work".

Re:Just as good (?)

[phorm]

You are missing a little bit. There's a reason why AMD started going with giving their processors numbers instead of clock ratings, mainly because clock ratings are starting to mean less and less.

Clock ratio is only one of the things that is indicative of a CPU's power... you might want to consider comparing the 64 and 32 bit varieties of this chip somewhat akin to comparing a 1Ghz P4 Vs an AMD, or a 266Mhz P1 vs a P2-266.

Re:Just as good (?)

[Anonymous Coward]

Why would I want to upgrade yet?

I don't know enough about you to answer that, but I'll assume that since you don't know why, there isn't a reason for you to upgrade, or you would know.

Where I work, memory space is the reason for our high end 64 bit machines. The problems they solve (3D seismic migrations) require random access to very large data sets (>4GB).

There are other problems (not here) that can be solved much faster with the more and larger registers.

Will this help your FPS frame rates? Not enough to bother.

Video encoding?

[Lussarn]

Have anyone tried to encode xvid with one of these in 32 and 64 bit, preferebly using Linux? Is there much difference in speed? I'm looking at the 3000+ part as it is cheap but there are zero and none benchmarks to back it up in 64 bit mode.

Re:Video encoding?

[satterth]

Could one assume that Xvid benchmarks would scale along aside the DivX benchmarks that are included with the artical? I know its not the same but it would be a good place to start.

Re:Video encoding?

[Anonymous Coward]

There are no 64bit compiled video encoders in the article.

Anandtech also has a review...

[Fallen Kell]

They mainly focused on the price/performance ratio as it is truely a killer to everything out there.

Link to Anandtech Article

Basically, they are predicting the death of the AMD FX51 as the 3400+ has equal or better performance and MUCH cheaper.

64-bit Linux on AMD-x64 reviews.

[eddy]

Please add links to any reviews that run 64-bit linux (or other 64-bit OS of choice) with 64-bit benchmarks on said processor:

fineprint: I don't need a lecture on the nature of 64-bitness.

We have all of the reviews listed

[ruiner5000]

Guys we have all the reviews listed on our main page, and I'm adding more as they come in. It currently totals at 19. Does Hothardware pay Slashdot for these links? ;)

Re:We have all of the reviews listed

[DAldredge]

Damn, you posted something that doesn't bash intel and/or apple.

I am in complete and total shock.

For those that do not know, AMDZONE used to be a great site for info on AMD. In the past 3-6 months it has become a bash Intel/Apple site.

Note. I am typing this from a dual athlon, and most of the other systems in my lab are AMD based. So do not call me an Intel/Apple fanboy.e

Not a very good review....

[imsabbel]

Well, that review sucked, at least be normal anandtech standarts. Hey, its a CPU bench, so why are 2/3 of the game benches gpu-limited and all processors are within 2-3%?

Fair Comparisons?

[Aaron England]

If they are comparing the $700 AMD 64-FX chip, they should be comparing it to Pentium's $1000 P4 3.2 EE chip, not their sub-$400 P4 3.2.

Also does anyone have an idea how expensive the AMD 3400+ chips are? Because the AMD 3200+ chips are $400 retail. The article quoted a price for a thousand quantities but I was wondering how much it would cost for just one. Because if its pricey enough the P4 3.2 may beat out the 3400+ dollar for dollar.

Though Intel doesn't have to really worry about that title. At $164 the Pentium P4C smokes the pants off any AMD processor in its price range. At least, after overclocking it to 3 GHz, which is very doable even with standard cooling.

Re:Fair Comparisons?

[ViolentGreen]

Check out the anandtech article from several posts above.

It's compare there to the EE chip. I believe that I remember hearing that the price for the 3400 is less then the FX chip and out performs it as well.

Re:Fair Comparisons?

[EmagGeek]

"Though Intel doesn't have to really worry about that title. At $164 the Pentium P4C smokes the pants off any AMD processor in its price range. At least, after overclocking it to 3 GHz, which is very doable even with standard cooling."

Will it really be cheaper and faster when you have to buy a new one every 6-12 months because you destroy it?

Re:Fair Comparisons?

[drfishy]

I'm not sure where you get your benchmarks, but for $144 you can get a Althon XP 2800+ from newegg that will smoke the pants off a 2.4Ghz P4 and overclock just as well... BTW, retail 3400s are also at newegg for $439.00... Personally I'd spend half that and get 95% of the performance with a 3000, which is probably the best price/performace processor ever (and certainly at launch)...

Re:Fair Comparisons?

[Just Some Guy]

Will it really be cheaper and faster when you have to buy a new one every 6-12 months because you destroy it?

Indeed the word "overclock" has become my hardware review spam filter; it has a strong "cold fusion" connotation. If I drop $700 on a CPU, I will not be running it out of spec in any way. If I'm that hungry for speed, I'll build a cluster.

I can understand people wanting to overclock, say, a P-III 933 to see how far they can push it, but I just don't get the fanboy fascination with extreme cooling, adding a few megahertz, etc. Reading this stuff in a tech article is like finding an article on adding a whaletail to a ricer in Car&Driver - it just doesn't belong in a serious text.

Re:Fair Comparisons?

[Anonymous Coward]

At $164 the Pentium P4C smokes the pants off any AMD processor in its price range. At least, after overclocking it to 3 GHz, which is very doable even with standard cooling.

And your benchmarks are where?? I for one am running an AMD Barton 2500+ overclocked to 3000+ (that's 1.8ghz OCED to 2.2GHZ for you people who still believe the mhz myth) which costs 90$ at newegg.com My 3dmark scores & pifast time are BETTER than others using the processor you have linked to. And it will run just as good or better in the things most people do id est, gaming, video watching, web surfing, business work. Sure you will get 5-10 FPS faster in video encoding, but it is a very very small minority of PC users who encode video more than 5 minutes in length. For the 74$ saved I can almost afford an 80 gig hard drive or one of the 120 gig if it's on sale.

Re:Fair Comparisons?

[keyne9]

Prices dropped today. The 3000+ is now approximately 220$, 3200+ is around 280$, and the 3400+ is around 440$. Newegg prices, of course.

Seems like the Anandtech review places it fairly highly as an FX51 killer =P

Re:Fair Comparisons?

[devnullify]

Yeah, except that unlike a whaletail, you actually get a real performance increase.

Personally, I'd rather have a stable 2.2GHz CPU than a stable 1.83GHz CPU for the same price (my 2500+). No extra cooling, .1v overvoltage, and ~58C temps under load. It's solid, I haven't had any lockups or other problems.

I don't understand the aversion to overclocking. I've got an old Celeron 633 that's been powering my server for about 3 years now 24/7 and it's still going strong. The CPU cost me $75 at the time I bought it. If it dies, I'll just replace it. Who cares that a $75 part lasted 4 years instead of 10, replacing it would be a drop in the bucket compared to the cost of purchasing a comparable performing CPU at the outset, and you'll definitely get a better performing replacement than you would've had if you spent the extra cash in the beginning.

By the time you're in any danger of causing real damage (like after 4-5 years of continuous use), you're going to want a new CPU anyways, so what's the problem?

Re:Fair Comparisons?

[Nazmun]

BTW, my cooling wasn't extreme... just a very large heatsink with good a large surface area and a 90mm fan (the larger the fan the lesser the rpms needed to move air and the lower the noise). Simply air cooling not water, peltier, or phase change... 50 % faster about 15 degrees cooler then most processors running on spec (the heatsinks amd ships are extremely weak... this is one thing intel has them totally beat at) as far as amd goes.

I'd just like to point out...

[222]

Any peice of hardware that can spank the competition EVEN while its potential isn't fully being realized by the software testing it deserves my dollar.
And yes, im talking about how well it games, I can really give a flying fsck about how quickly it runs office...

64 bit

[Ba3r]

The athlon gets stomped in alot of the benchmarks here, but it seems there is little out there to take full advantage of the 64bit architecture. Especially when it comes to memory addressing, and I am sure there are people out there looking for more the 3 gigs of ram.

Re:64 bit

[Dravik]

If you read the article it mentions that most of the benchmarks have the AMD 3400+ beating the P4. There are a bunch of additional tests that are not quite the same as benchmarks. The difference is mentioned in the conclusions at the end of the article.

OMG ROFLMAROESOREO

[Anonymous Coward]

YUO CALLED IT 'ITANIC' BECASUE IT SI A DISASTRE LEIK TEH TITANIC BECUASUE IT SI A SHIP TAHT SANKED!!1!!~~!tilde OMG ROFFLE YUO SI TEH +5 FUNNY OMG OMG OMG ROFFFLFLFELELFLEFL YUO ARE TEH AMD FANBOI BECAUSE AMD SI TEH ROXXXXOORRRSSS!!!

Due to excessive bad posting from this IP or Subnet, comment posting has temporarily been disabled. If it's you, consider this a chance to sit in the timeout corner. If it's someone else, this is a chance to hunt them down. If you think this is unfair, please email moderation@slashdot.org with your MD5'd IPID and SubnetID. Due to excessive bad posting from this IP or Subnet, comment posting has temporarily been disabled. If it's you, consider this a chance to sit in the timeout corner. If it's someone else, this is a chance to hunt them down. If you think this is unfair, please email moderation@slashdot.org with your MD5'd IPID and SubnetID. Due to excessive bad posting from this IP or Subnet, comment posting has temporarily been disabled. If it's you, consider this a chance to sit in the timeout corner. If it's someone else, this is a chance to hunt them down. If you think this is unfair, please email moderation@slashdot.org with your MD5'd IPID and SubnetID. Due to excessive bad posting from this IP or Subnet, comment posting has temporarily been disabled. If it's you, consider this a chance to sit in the timeout corner. If it's someone else, this is a chance to hunt them down. If you think this is unfair, please email moderation@slashdot.org with your MD5'd IPID and SubnetID. Due to excessive bad posting from this IP or Subnet, comment posting has temporarily been disabled. If it's you, consider this a chance to sit in the timeout corner. If it's someone else, this is a chance to hunt them down. If you think this is unfair, please email moderation@slashdot.org with your MD5'd IPID and SubnetID.

Re:Obligatory Tolkien Reference

[Anonymous Coward]

I see an analogy between AMD/Intel and Middle Earth.

When was the last time you were outside?

The game of speed

[strider69666]

You know, I myself just bought a 2500+ on the 4th for $80.00 and couldn't be happier with it. It is overclocked from 1.8Ghz to 2.1Ghz with a standard Volcano 7 heatsink and fan. This new 3400+ sounds like a fast chip, and 64 bit is the shiznit. BUT, it will probably cost $400 on release, and stay above $200 for at least a year. What I am getting at is that although chip makers obviously need to make their money, wouldn't they make more by using the best business model: sell more for less. I mean, instead of (just for demonstrative purposes) selling the 3400+ at an introductory price of $400 and selling maybe 100,000 they could sell it at an introductory price of $300 and sell 300,000 and make up the price difference in the volume, plus some. I used to work at a friends video game store, and when the X-BOX came out, instead of selling them for $299.00 and making $20.00 per box, and selling maybe 3 a day, he was selling them for $289 and selling out every day, with like 10 sales a day. It is a good example of an incremental decrease in price exponentially increasing sales/profit. Especially when the market is full of other products that are so close in feature/speed/compatibility for much less cost to the consumer.

Re:The game of speed

[edwdig]

You're forgetting that AMD has a variety of different speed chips out on the market. Some people just want the fastest chip they can buy, and will pay whatever it costs to get it. Those people are the reason chips cost as much as they do when they come out.

Most people however simply have a rough estimate of what they want to spend on a computer, and buy the best they can afford for that price. They're just as happy with a 3000 or 3200 instead of a 3400.

The current pricing scheme works fine - get big money from the people willing to pay anything for the best there is and let everyone else pay a lot less for a processor not that much slower.

Re:The game of speed

[adolf]

Your X-Box-selling friend was making 66% more money daily, but was expending more than 3 times as much work, shelf space, and time to do so.

It is a good example of diminishing returns, not of some "incremental decrease in price exponentially increasing sales/profit."

Remember kids, overhead costs money.

Re:The game of speed

[MoralHazard]

For starters, your friend was undercutting every other X-Box retailer in the country by doing this. He wasn't selling 10 per day because the price was lower--he was selling them because they were cheaper than everybody else's price. Imagine if most retailers had followed his lead and lowered their price to $289 a box, also. He would have been selling the standard 3 per day, same as everybody else. So no, this ISN'T a good example of what you were trying to prove.

AMD is in a somewhat different situation because they sell ALL of the Athlon CPUs that are bought. They can't undercut anyone else to offer the same product for less money. Sure, you could argue that Intel's chips are similar to AMD's, but I think they're different enough. An Athlon64 is neither A) a simple replacement for an Intel P4 (because you have to buy a new mobo, at the very least), nor does it B) provide the same type of benefit to the consumer (the two types of CPUs perform differently on various loads).

Since a consumer can't automatically substitute an AMD chip for an Intel chip, AMD would have to drop their prices low enough to overcome the cost and hassle of the consumer buying a new motherboard to fit the other chip, unless the consumer is shopping for a board and chip at the same time (some do, some don't). Also, AMD would have to drop their prices low enough to overcome the perception in the market that their chip isn't as fast/good/special as the P4. And keep in mind, when your product is cheaper than the competition, people wonder WHY it's cheaper, and generally assume that it's because your product isn't as nice--"you get what you pay for" is true often enough.

Those X-Boxes are all exactly alike, unlike the P4 vs. Athlon64 argument--since there are differences between Athlon64 and the P4, a SIGNIFICANT price difference would have to exist for the effect you describe to exist.

Re:extra links

[Anonymous Coward]

nice try, Steve... look at the user number. Eyegor's been around a while

So?

[eddy]

Did you fix the Lost Circuits-link yet?

Tilda vs. minus

[crow]

I expect they meant to use a tilda ('~') instead of a minus ('-'), so as to indicate "about" instead of "negative."

The best a heatsink can ever hope for is to cool to the ambient air temperature, and we won't see anything aproach that until we have superconducting heatsinks. (Imagine a large superconducting mass in the ground with a superconducting cable connecting it to the CPU to draw off heat: power outlets with a pin for cooling, superconducting traces on circuit boards for cooling, and no need for fans.)

Re:Tilda vs. minus

[Anonymous Coward]

More likely he intended to say "mid-40C" instead of "mid -40C".

Re:Tilda vs. minus

[Anonymous Coward]

Of course if you have a good enough fan it will suck a vacuum before the sink which absorbs faster particles easier leaving slower ones there thus cooling it further than air temperature would allow.. ;)

Re:Tilda vs. minus

[chalco]

He may have meant a ~, but I doubt it. You want to google something called phase change cooling. It is available in commercial products such as the Prometia or the Vapochill. Phase change cooling does achieve temperatures that low. I was shocked when I first learned about it, thinking it was a typo myself. The technology is not a heatsink, it is more like a radiator.

Re:Tilda vs. minus

[niko9]

I'm not trolling, just trying to educate.

It's called a tilde:

nick@marvin:~$ dict tilde
3 definitions found

From Webster's Revised Unabridged Dictionary (1913) [web1913]:

Tilde \Til"de\, n. [Sp., fr. L. titulus a superscription, title,
token, sign. See {Title}, n.]
The accentual mark placed over n, and sometimes over l, in
Spanish words [thus, [~n], [~l]], indicating that, in
pronunciation, the sound of the following vowel is to be
preceded by that of the initial, or consonantal, y.

From WordNet (r) 2.0 [wn]:

tilde
n : a diacritical mark (~) placed over the letter n in Spanish
to indicate a palatal nasal sound or over a vowel in
Portuguese to indicate nasalization

From The Free On-line Dictionary of Computing (27 SEP 03) [foldoc]:

tilde

"~" {ASCII} character 126.

Common names are: {ITU-T}: tilde; squiggle; {twiddle}; not.
Rare: approx; wiggle; {swung dash}; enyay; {INTERCAL}: sqiggle
(sic).

Used as {C}'s prefix {bitwise negation} {operator}; and in
{Unix} {csh}, {GNU Emacs}, and elsewhere, to stand for the
current user's {home directory}, or, when prefixed to a {login
name}, for the given user's home directory.

The "swung dash" or "approximation" sign is not quite the same
as {tilde} in typeset material but the ASCII tilde serves for
both (compare {angle brackets}).

[Has anyone else heard this called "tidal" (as in wave)?]

(1996-10-18)

Re:Tilda vs. minus

[tiger99]

It is not uncommon to use a Peltier Effect cooler. This is basically a huge stack of thermocouples run in reverse. You put in a lot of current at low voltage, and it produces a temperature differential. If you heatsink the "hot" side to ambient air, the "cold" side may be well below zero. But, it is not very efficient (like all cooling systems), so you need to put in several times as many watts as it extracts from its "cold" side, and not surprisingly the total of both appears at the "hot" side, so you may need a very big heatsink with powerful fans.

I don't have the numbers in front of me right now, but at a guess you would need 300 watts to cool a 100 watt CPU, so would need to dissiapte 400 watts to air.

It is inadvisable to make any attempt to get the chip below zero, obviously ice formation will happen, and when you switch off, it will melt. Should you switch on again, disaster is quite probable, unless the PCB had a good conformal coating and the socket has an interfacial seal. The conformal coating can be dealt with quite easily, but I have never seen a sealed CPU socket. BTW I usually work as an avionics designer, where we have to make things that will run from well below zero to well above, so I do know the problems.

Another issue is thermal fatigue. The temperature coefficient of expansion of silicon does not exactly match that of the (probably epoxy) package, every temperature cycle causes a stress cycle, which causes a strain cycle, until something breaks. Same for the motherboard itself of course, if you should cool the whole thing. That is also a good reason to never overclock anything, apart from the possibility of getting subtle data errors and increasingly buggy OS as a result of inadequate timing margins, you will definitely wear the thing out a lot quicker. Every 10 deg C roughly halves the life, or the number of on/off cycles it will survive, and if you do the calculations, the numbers are quite depressing for a modern PC.

If you really want a thumping great 64 bit processor (I certainly do, when the price comes down!), it would be best to calculate the cooling system, and maybe do some tests with thermocouples etc, to try to get the CPU chip to settle down at a relatively safe temperature, say 40 deg C, without getting ice formation on the coldest parts. The clever bit would be to get it to power on and off without any excursions below room temperature (often 20 deg C) or above 40 deg C. Heat soak when you switch off the CPU would be minimal, the mass of the chip itself is very small, but cold soak from a huge peltier block could be a problem, the CPU could be dragged down to -40 deg when you switch off, which is exactly what you don't need, for a long and reliable life.

The other thing to watch out for is that at low temperature the CPU internals will be out of spec. It is actually possible to get excessive current flow in some transistors, and local hot spots, because it is too cold. There may also be timing problems, data corruption, ...... It is not possible to test properly that these things are not happening. It takes a smallish time to fully exercise an 8-bit processor to verify all possible data and instruction operations, but somewhere near the lifetime of the universe to do the same for 16 bits. Throw in onboard cache, 64 bits, etc, and it is just impossible. These things work statistically, AMD know how timing variations, for example, might vary across the chip, and allow sufficient margin, within the published clock frequency and temperature range, but deviate from these ratings and this is no longer true. If running at 1GHz, a 1 in 10e12 error rate would corrupt your data or OS code within 1000 seconds, just over 1/4 hour. The error rate required to run an OS for weeks at a time at several GHz defies all attempts at testing. Again, a very good deterrent to overclocking (BTW my non-overclocked ancient K6-350 had a meltdown due to fan failure, and as it died the corrupt Win XP blew away all the passwords so when I got a new CPU and fan, any of th

Re:Tilda vs. minus

[stanmann]

What is an on-off cycle. Is that like rebooting. Or is that when you turn the computer off for upgrading... Or are there still people who think that the power savings is worth the wear on the hard-drive bearings?

Re:Tilda vs. minus

[Anonymous Coward]

Superconductors are actually very poor conductors of heat - the paired electrons don't contribute to the thermal conductivity and in a typical metal the electron thermal conductivity is a significant fraction of the total conductivity

they are perfect electrical conductors, but poor thermal conductors

Re:Tilda vs. minus

[crow]

Hmmm. I guess I can't believe everything I read in science fiction books.

Re:Tilda vs. minus

[Anonymous Coward]

well, I wouldn't take everything as fact!

But it's not inconceivable, there are also superfluids, which flow without viscosity. Imagining some sort of thermal superconductor seems not totally unreasonable.

What's the point without 64-bit OS and apps?

[Masarand]

The main point of the Athlon 64 and Opteron is that they are 64-bit CPUs that can run 64-bit applications.

The fact that they can run 32-bit apps under a 32-bit OS at pretty much the same speed as a 32-bit CPU is surely a huge yawn (but great for backward compatability.)

Has anyone seen any comparative benchmarks under a 64-bit Linux system?

Re:What's the point without 64-bit OS and apps?

[lederhosen]

They had some on aceshardware.com for a while ago. The Athlon64 did better when compiled to a 64 bit binary. Most of the gain is that you get more registers, not that you can handle 64 bit registers. They got some more speed when using the NUMA support in the kernel (Some of the tests where SMP tests).

Re:What's the point without 64-bit OS and apps?

[Anonymous Coward]

The point is they will make money off of the morons who need to have the latest and greatest technology. So they can brag about what new toys they purchased. When these chips drop $800-$1000 in price a year or so from release people like you and I can buy them relatively cheap.

Mobo selection

[ackthpt]

I'm in the market for a new Mobo/CPU, to upgrade from an XP 2600/333. It looks good and appears to demand a bit less power, however, after all the trouble I've had with my present mobo (Asus A7V8X) I'm still iffy on dropping the cash. Processors always look good, mobos (now that I'm a bit cynical) all look like dressed up used cars you don't want to look under the hood of.

Recommendations on a good solid board for one of these? (I don't have money to go out and buy new boards and stack them up as dust collectors, probably same as most, but some people appear to have an endless supply of cash to play with until they get their boxen right.) Are Asus SK8N/SK8V any good, or problematic? I basically what to put this thing together, slap in about 8G of RAM and get back to work, thanks in advance.

Re:Mobo selection

[TheCoop1984]

You're upgrading from that??? My 2500+/333MHz works perfectly for what I want to do, I won't be upgrading it for a year or two at least...

Re:Mobo selection

[ackthpt]

Start video editing and see how you feel about it.

<zombie>More memory, must have more memory... </zombie>

Re:Mobo selection

[Anonymous Coward]

Before getting my amd64 I did some research to find that the VIA chipset is the best right now. nVidia is apparently too imature. Therefore I bought a GigaByte K8Triton from newegg.com where I also picked up the cheaper 64 3000+, which is same as 3200+ but for the cache size; this is a big deal but saved me 200.

NR

Re:Mobo selection

[Anonymous Coward]

All of the reviews I've seen put the 4 current chipsets (nVidia nForce 3/SiS 755/ALi M1687/VIA K8T800) within 3-4 % of each other, with the VIA K8T800 *just* in front.

At the end of the day it really comes down to the other features on the motherboard, such as Gbit ethernet, SATA contollers.

Toms Hardware (I know how much /. loves it) has a current review of these boards with manufacturers details on HyperTransport/HyperStream/V-Link comparisons (which look to be manufacturer supplied) which is worth having a look at.

Re:Mobo selection

[ncc74656]

Start video editing and see how you feel about it.

<zombie>More memory, must have more memory... </zombie>

IME, video editing/compression isn't much of a memory hog. Compression, though, will chew up all the processing time you can throw at it. A faster processor (or multiple processors) will do you more good than a ton of memory.

Re:Obligatory Tolkien Reference

[MagicMerlin]

When was the last time you were outside?

Well, my comment was meant to be taken strictly tongue in cheek. Evidently the moderators take these things pretty seriously. What in the world was I doing, thinking about hobbits?

I claim underrated :)

Merlin

TechReport review

[JohnnyBigodes]

You may also want to take a look at this review at our good old Tech Report.

Why do speedometers go above 100MPH?

[dpilot]

On most of the roads in the nation, the speed limit is either 55MPH or 65MPH. Some places out West on the Interstates, it's 75MPH. Even a 100MPH speedometer is WAY overdesigned, well past short-term bursts for passing, accident avoidance, and the like.

So why do we have speedometers that go up so high, and why can many cars actually go that fast? After all, it's illegal, and we don't NEED that speed, or speedometer.

Perhaps we really do - about as much as a 64-bit processor.

Re:Why do speedometers go above 100MPH?

[0123456]

I take it you're not old enough to remember the stupid 85mph speedometers which used to be fitted to most (all?) American cars?

Re:Why do speedometers go above 100MPH?

[dokutake]

Because cars are geared to go the speed limit at low revs in top gear for fuel economy. When you decide to screw fuel economy, then you can obviously go a lot faster.

Re:Why do speedometers go above 100MPH?

[ViolentGreen]

That's the reason I drive a Nissan and not a Porche. Well one of them at least...

I can see the relation but I'm not sure of your point.

Re:Why do speedometers go above 100MPH?

[AKnightCowboy]

So why do we have speedometers that go up so high, and why can many cars actually go that fast? After all, it's illegal, and we don't NEED that speed, or speedometer.

I don't know about you, but if I'm only doing 100mph you're probably in front of me in the fast lane slowing me down. Get the hell out of the way so I can pass without swerving in and out of traffic.

Re:Why do speedometers go above 100MPH?

[Anonymous Coward]

I don't get what you are trying to say. Is a 64 bit processor an unneeded danger?

My car goes 130 MPH. Some day I will take it on a track and see what it's like to drive over 100. I like a car that can go that fast because I do a lot of 75 MPH driving. I've had cars that top out at about 80 and they are scary at 75. With my current car, I have great handling and lot of power at 75.

We use 64 bit processors here at work on machines that have more memory than a 32 bit processor allows. Lets say current systems have 512 MB and memory size doubles every 18 months. In three years most new computer will require 64 bit processors or be short on memory. Of course you can always just pull out the "2-4GB" should be enough for anyone line.

Re:Why do speedometers go above 100MPH?

[addaon]

Ugh. There used to be a law in the US that speedometers had to top off at 85 mph. My car ('93 Explorer) has that, and it's damnably unsafe. Most of the time, I have no fucking clue how fast I'm going, other than "as fast as the cars around me." I've taken to using a GPS on I-280, just to know if I'm going under 100 (at which point the cops might care, since I'm obstructing traffic going so slow) or not.

Re:Why do speedometers go above 100MPH?

[VVrath]

Thinking like that that was a contributory factor in causing Apollo 13's oxygen tank to go kaboom.

When voiding the tank after a test, a faulty thermostat left the heater used to boil of the oxygen permanently switched on. This would have been noticed if the teperature gauge inside the tank didn't top out at the high end of expected 'normal' operating temperatures.

If your speedo topped out at 55/65/75MPH (for example) you'd never know for certain if you were travelling at that speed, or above it.

Re:Why do speedometers go above 100MPH?

[dpilot]

I can see I need a bulk response here. You get it, for everyone.

My point wasn't to be taken too literally. I was trying to say that most (but not all) of us will never drive over 100MPH. Most (but not all) of us have no need for a speedometer that goes to 120 or 160 MPH, but we all have them. That's not to say that our speedometers shouldn't have some margin above the top US speed limit of 75MPH, to debunk the Apollo analogy, a little. We need some margin, just not 100% margin. (It may be that Montana is back to 'reasonable and proper' for speed limit, but I'm not sure. Last time I was there, Utah and Colorado were 75MPH.)

But we'd all feel 'cheated' if our speedometers capped off at 85 or 90 MPH. We wouldn't like Detroit saying "You don't need that capablity." For that matter, most of the capabilities of BMWs or Porsches, or for that matter, SUVs and Hummers are wasted on our roads and highways. But people want them, and buy them, and spend WAY more money than needed for 'basic transportation' on them.

The same is true of CPUs. Most (but not all) of us don't need 64-bits on the desktop. Most (but not all) of us don't need a fraction of the 32-bit capability that we've got.

But we're all buying bigger, better, faster - anyway. Needed or not, 64-bit CPUs will sell. Once they have risen to the consciousness of Mr/Ms BMW/Porsche/Hummer, and once those people can make the *frightening* choice to buy non-Intel (Unless Yamhill is marketed in time) CPUs, they'll believe they NEED a 64-bit CPU.

Re:Why do speedometers go above 100MPH?

[Anonymous Coward]

Because Some people do not live in America! Yay - In Nigeria, you can drive as fast as you can afford - you don't need a speedometer if you have enough Naira!

Re:Why do speedometers go above 100MPH?

[Anonymous Coward]

Montana had no speed limit on interstates for a couple years, also. I think that was only during daylight hours, though.

AMD's chance

[G3ckoG33k]

I wouldn't be too surprised if AMD chose to withhold faster versions of the Athlon64 FX-series until any Prescott is just about to be released. A day before or so. Leap-frogging at its finest.

Re:I just want to say

[Bert64]

that all you fucking Intel fanboys can suck my third testicle. If you're too cheap to buy CRAY, then you're too cheap to buy good memory and fast disk arrays, so you whine and complain when your crappy hardware craps out on you.
My Cray X1 with 4096 CPUs and 65536 gig of RAM will stomp on anything you cheap bastards can build with Intel shit. You think your so hot with your 32bit 3ghz processor when you dont realise that any system with less than 128 cpus is just a joke.

Forgive me, I live with english majors.

[xankar]

Unlike the Athlon 64 FX-51, this new 3400+ rated Processor, has a 64 bit memory interface...

Commas, are awesome. Especially, when used in excess.

Pentium Vs. P4

[wowbagger]

Well, in all my testing of running 16 bit apps, a Pentium I outran a similarly clocked P4 by a healthy margin - so obviously the Pentium is a better chip, right?
</sarcasm>

Seriously - For a period of time the A64 will be running mostly 32 bit apps (at least in the Windows world), and so it is fair to benchmark its performance against 32 bit apps. But I cannot help but wonder how much P4 tweaking all those apps had, and how much A64 tweaking they did not have.

Also, the memory performance tests are, to my mind, somewhat questionable as well, as different CPUs even within the Pentium line have different memory access behavior - code that will be bus limited on a P4 might not be bus limited on a P3.

I am not saying the comparisons are not useful, but I am saying that they don't tell the whole story. Let us see some benchmarks wherein the A64 is running code that is written for the A64 - using the extra registers and so on.

Punctuation is your friend...

[HBK-4G]

...but you abused that submission with more commas than a Microsoft EULA has in total. Learn to use that 'preview' button for more than just HTML-checking.

Misunderstanding Thermodynamics

[EmagGeek]

There is no such thing as a thermal superconductor, at least not that I am aware of. There are electrical superconductors, but everything has a specific heat which limits the rate at which it can conduct heat. Please correct me if I am wrong.

Also, the best a heatsink can ever hope for is to cool to some temperature that is pretty far above ambient. You would need the capacity to dissipate energy to infinite space in infantessimal time in order to cool an energy dissipating device to ambient temperature, without getting active that is.

The best a heatsink can home for is the product of the energy being dissipated and the total thermal impedance of the package-to-heatsink-to-air interface... Increasing air velocity across the sink will only help so much as the energy added to the sink from compression of the air falling on it will eventually surpass the amount of energy taken away by it...

Re:Misunderstanding Thermodynamics

[imsabbel]

There is the phenomenon called superfluidity, which is closely connected to "thermal superconductivity".
Helium4 below 2.6 or so Kelvin becomes a superfluid. In this state, there cant be a thermal gradient anymore. Heat is transfered with the speed of sound.
If you would use this to cool a cpu (doesnt work, because the doted SI becomes not so dotted at that temp), there wouldnt be any bubbles, only evaporation on the surface.

Even if its,e.g, a long heatpipe.

Re:Misunderstanding Thermodynamics

[EmagGeek]

Shit, and how much energy are you going to expend keeping Helium4 that cold, along the entire length of the pipe? :)

Re:Misunderstanding Thermodynamics

[imsabbel]

Easy. Just keep the presure at around 30-40 mBar. Pump away any gas that evaporates. The at long as the pressure doesnt increase the rest will keep its temprature.
Of coure you need a lot of helium, some Really Big turbo pumps, and a LOT of energy (around 50 to 100 times the amount the helium absorbs...),ect... (i didnt say it was trivial...)

Oh no they won't.

[mapnjd]

Most apps don't need to be 64-bit, it doesn't make them faster, kewler or anything.

For example Solaris 8 only comes with 32 64-bit binaries (/usr/bin/sparcv9) and they're almost all ones which look at the process table.

Re:I just want to say

[Anonymous Coward]

Wow, crays are fast dude. Imagine a beowulf cluster of those!

"Ignorance" gains a connotation...

[Anonymous Coward]

Which in this case is "wilfully ignorant". Given modern usage, there is a difference between saying "person X is uninformed and asking" and saying "person X is ignorant". The contention over the use of the word "ignorant" is related more to the unstated but clear insulting connotation that the original poster was "wilfully ignorant", with which I disagree.

Usually posting as "lenski" but too lazy to login...

Java VM is exactly what we don't need

[arth1]

The Itanium is too expensive and slow. Ditto Sparc. AMD 64 bit servers running 64bit Java VMs will make for a killer combination.

First you think taht Itanium and Sparc are too slow. You want speed -- good!
But then you ruin it all by running java and 64 bit[1]?

Java was never meant for speed - it was meant for compatibility, ease and safety of programming. If speed is the most important factor, you choose something else.

Also note that 64-bit is not faster than 32-bit in and of itself -- it is faster when larger values than what fits in 32 bits is needed, or you need to address larger parts of memory.
If you don't NEED values exceeding 32 bits (most java apps will fall here), 32 bits is almost always going to be faster -- if nothing else, because less data has to be transferred (and cached) from RAM to the CPU. This is why IRIX, for example, lets you run their 64-bit chips in 32-bit and mixed 32/64-bit modes.

Regards,
--
*Art

Re:Java VM is exactly what we don't need

[Juergen Kreileder]

Also note that 64-bit is not faster than 32-bit in and of itself

It is on AMD64!

I've ported the HotSpot VM to AMD64 for Blackdown. It's noticable faster the 32-bit version in allmost all benchmarks. The main reason for the performance gain is that you have more registers in 64-bit mode.

Re:Java VM is exactly what we don't need

[Master Bait]

A two-lane road is just as fast as a four-lane freeway as long as there are only a few cars on the road. But the reality is the roads are quite crowded these days

Re:Java VM is exactly what we don't need

[EvilBudMan]

Also the memory controller is onboard. I just built 2 2.0ghz machines with an MSI KT8 neo mother board and 2 WD SATA drives in raid 0 and am more than satisfied with the result.

Re:Java VM is exactly what we don't need

[arth1]

A two-lane road is just as fast as a four-lane freeway as long as there are only a few cars on the road. But the reality is the roads are quite crowded these days

Your analogy is seriously flawed. The CPU isn't the road, it's the CAR.
Going from 32-bit to 64-bit is like making the cars twice the size. This might help, as you can transport more in a bigger car, but if the amount of data transported isn't the bottleneck, it's not going to help much. In congested situations, it will actually HURT performance, and it won't be as useful as its potential until the whole infrastructure is capable of catering to the bigger cars.

Going from two lanes to four lanes would be like doubling the maximum bus width without increasing the bus width per CPU.

--
*Art

Gad, another standard review.

[Jeremy Erwin]

The reviews are all the same--run various permutations of the PC through benchmarks, and display the results using bar charts. And not just any bar charts. Use a gradient to color the bar, so that the color legend is rendered useless.

The reviewers should read Tufte, and figure out a more effective way of illustrating their analyses than endless pages of bar charts. Oh wait, that's how they get their ad revenue. Never mind.

I hate these stupid reviews

[Schugy]

They always use the same outdated games and a lot of synthetic benchmarks. Much more important is that they dont use a SuSE9 for AMD64 and they dont tell me whether the 512kb extra cache are still unimportant in 64bit mode. Maybe the 3000+ sounds like the better choice but fails in 64bit mode compared to the 3200+.

need? no. demand? yes.

[mapmaker]

I still don't think there is a huge demand to have these in desktops as of yet.

There is a HUGE demand for these desktop chips. AMD has pretty much sold out of them.

Your point about people not really needing these processors is valid, but the demand is there.

AMD prices:

[Kjella]

In lots of 1000:
Athlon 64 FX-51: $733
Athlon 64 3400+: $417
Athlon 64 3200+: $278

So retail using your $400/3200+ as mark-up ratio, that should be something like $1040, $600 and $400, respectively.

Also, on the P4/Athlon war I haven't checked lately since I'm happy with the XP2000+ I have, but at the price range I've been at AMD has come out on top for my last three processors (Duron 700, Athlon 1200 and the above mentioned XP2000+). Maybe the P4C is different, right now I really don't care though :)

Kjella

Athlon64 3400 review on Tom's Hardware

[glinden]

Tom's Hardware has an interesting recent article on this chip too. Lots of benchmarks.

Hush

[steveha]

won't a P4 system give me better "speed per dollar"?

Actually, an AthlonXP will give the best speed per dollar, since it gets more done in a clock cycle. It's actually pretty close between Athlon and Pentium, but if you add in the cost of the electricity over the life of the computer, the AthlonXP will win.

What I would care about more is a silent and small (think book sized) system. When I say _silent_ (not just almost silent), I mean that it won't need a CPU fan, no power source fan and that it would be based around a 1GB compact flash card.

I have to quibble about the CompactFlash card. It really isn't a very good way to set up your computer: an OS that expects a hard drive will wear out your CompactFlash quickly. If you are willing to run Linux, and you don't put /var or /tmp on the CompactFlash, this could work. (Make /usr and /etc read-only, and maybe make a CRAMFS for /home or something.) But you would be best off just accepting a hard disk; the hard disk is not the source of very much noise. Get one with fluid bearings. For complete silence, just do a net boot, and run your system purely out of RAM.

Note that you really don't need a 1 Gbit ethernet connection; 100 Mbit will be just fine for a net boot config. I used to use UNIX workstations that had NFS mounted directories over 10 Mbit Ethernet and I rarely noticed any speed difference from the local directories. Not that many years ago, a 10 MB/sec drive interface was considered "fast", and 100 Mbit/sec is about as fast as that. I think you will notice a busy CPU more than you will notice the network speed.

As for the rest of what you want, it is possible with a Via C3 today. It will be possible with a Crusoe chip when the Mini-ITX Crusoe boards ship.

I have a serious case of techno-lust for a Hush computer. Look how well it meets your specs, right down to a fanless power supply (it has a transformer "brick" in the power cord). Note that they offer Seagate drives with fluid bearings (quiet) and even 2.5" laptop drives (should be even quieter, but also slower).

http://mini-itx.com/store/hush.asp?currency=2

Directron doesn't carry it, but they might someday. If you live in the USA, let Directron know you want one, please! If you order from mini-itx.com, you need to pay for shipping from England.

http://store.yahoo.com/directron/hushminiitx.html

Check out the other Via C3 options. The Tranquil PC also looks good:

http://mini-itx.com/store/tranquilpc.asp

I have fond memories of the Atari 520ST. Ours had two floppy drives, and no cooling fans at all. It was silent when you weren't typing. I'd like an email/web/writing computer that is silent like that.

steveha

Re:Hush

[steveha]

I just noticed the Hush ATX system. It seems to be passively cooled, yet it has a Pentium 4. Presumably they could also build one with an AthlonXP but they don't seem to have one yet.

http://mini-itx.com/store/hush-atx.asp

steveha

Thats Nice

[Nazmun]

You are obviously not a hardware enthusiast (this may sound like a flame but i just feel too strongly about this)... You can easily overclock for economical reasons without sacrificing cpu life. I've done this for years and continue to do so. As long as you know what your doing and have proper data from research (from reviews like this)you can take a sub $100 dollar processor and have it run like one for over $400 for years). The best part is that if you have the right components and you know how to tune your system you may even run more reliabley then systems running in spec with lower quality hardware (mobos, heatsinks, powersupplies, etc.)

Example:
I've overclocked processors by 50% (in terms of prices... a sub $100 dollar to a $500+) an example of such a processor that has been running for a year is by tbred b aiuhb core xp 2100 from 1700 or so mhz to 2400 mhz and a 66 mhz gain in fsb (which is huge as it's from 133-200 mhz). Note that no athlon xp that amd released ever surpassed the 2300 mhz level or the 200 mhz fsb (even the 3200 + barton) so that tells you much money i've saved. I bought in mid december of 2002.

The processor has been active for most of it's life either folding(when i'm not doing anything), gaming, or divx encoding.

Re:Thats Nice

[EmagGeek]

No, I'm not a hardware enthusiast.. I just design it for a living... There is a _reason_ specifications are written the way they are. You may not see the consequences of violating the specs, but they are there whether or not you want to admit it.

Granted, most people don't care if their game crashes more often or if they have to restart explorer every 10 minutes, but for people who use their computers to do useful things, taking chances is not the right thing to do...

Re:Thats Nice

[Nazmun]

It's nice that you design hardware but from the sound of it you don't work for intel, amd, or any of the relevant processor and chipset companies. Also you don't seem to be aware of many of these companies practices of speed binning processors (or whatever it's called nowadays) in which cores capable of higher speeds are packaged in as lower speed processors to satisfy demand of a certain speed/price range.

Next you talk about games and windows crashing... Well, all of my systems are thoroughly tested (individually too... most companies don't do this but they are safe as they don't push their machines to the max but often times that isn't the case) for periods of over 24hrs running things that greatly stress vital components whether they are overclocked or not to make sure everthing is working 100%.

The results speak for themselves as coupled with great programs like Norton Antivirus my Win XP machine has an uptime thats probably higher then you'd imagine. Last time i checked it reached 690 hours without a single restart which is pretty awesome despire xp's bad press and lord knows i push this system.

On the other hand my cousin has his athlon in his compaq system running about 20 degrees higher then mine... that in itself is not bad if not for the fact that his explorer does indeed crash every ten minutes.

Re:Thats Nice

[SaDan]

There's still no real point in overlocking anymore. A couple years ago, yes, but not now... Prices have dropped.

The time and money spent on a computer intended for overclocking is better spent on buying what you can afford, then upgrading when needed.

Yes, I used to overclock, A LOT, so I've been there, done that, know what you're talking about. Had plenty of super-stable, very fast systems, both Intel and AMD.

It's just not worth the effort these days.

Re:Thats Nice

[Nazmun]

Your probabyl right on the benefits but it's still interesting to do. Especially with some speed binned amds that can go up to 50% faster.

Re:Thats Nice

[EmagGeek]

Then again, there are hundreds of reasons that Explorer can crash every 10 minutes...

Yes, speed "binning" is a point, but it still doesn't justify the risk for customers who have critical applications running. I also question whether it even makes sense from a business standpoint. AMD would make more money by NOT downgrading processors because they would have more higher-end, expensive processors to sell. I guess it makes sense to take the loss if you're trying to eat up market share by offering as many cheap-ass CPUs as you can, but it hurts the people who need higher end chips by inflating their price...

You are also correct that I do not work for any of the chip makers, but the products I design are those that use their chips (embedded 603e cores mostly). We have to deal with margins and things of that nature all the time. Obviously, if we try to overclock to save money, things explode and we're liable :)

Re:Thats Nice

[Nazmun]

No one suggested thusfar to overclock in such a business/critical environment.

Does it run OS X?

[Anonymous Coward]

If not, I'll pass.

Alternative comparisons...?

[MrNemesis]

Maybe OT, but here goes...

I've been looking for ages for a decent roundup of how the Opteron series compares to the more mundane Athlon MP's and Xeon's (as well as Athlon XP's and P4's compared to the single processor Athlon 64's). I'm looking to build a new PC in the next 6 months or so (I do lots of video rendering, both in W2K and Linux) but haven't managed to find a comparison of, say, 2*244's vs. 2 Athlon 2800 MP's. I don't want to go splashing out on the very cute but expensive 64 bit chips if two dirt cheap Athlon MP's will do the job almost as well, particularly under a distro like 64 bit Gentoo. Preliminary resuts say that the Opterons and their ilk are blindingly fast but you can't blame me for being a stingy geek :^)

Any takers? I'm assuming there must be a few render-fanatics out there richer than I who've been able to do a comparison like this...?

These links below got me started, but IMO there's still an awful lot of ground that hasn't been covered - why oh why must EVERYONE use Windows in their benchmarks?!?!

http://www.3dchips.net/content/review.php?id=67& pa ge=1

http://www.anandtech.com/IT/showdoc.html?i=1935& p= 1

If anyone has any info or opinions, I'd be much obliged!

Re:Alternative comparisons...?

[Schugy]

I cant find any useful information on http://www.anandtech.com No graphs there in konqueror and opera 7.23

Re:Alternative comparisons...?

[PCBman!]

Have you tried Ace's HardwareThey have a server guide comparing Dual Xeons and Opterons along with a Quad Opteron system. You can also cross reference with their workstation tests for some dual Athlon MP tests.

I know they used Linux in their server guide review, but I'm foggy on their older reviews since I haven't read those for quite some time.

Re:Alternative comparisons...?

[MrNemesis]

Excellent!

I'd looked at Ace's but I didn't check their servers section, only the review.

At last, some relatively sane benchmarks! And no UT2003 in sight!

Thanks! /me starts saving for an Opteron

Flash on firebird.

[Anonymous Coward]

http://texturizer.net/firebird/faq.html#q2.1

Re: -40

[Anonymous Coward]

Or you could just say -40, since it's the same in Celsius or Fahrenheit :-)

Ok ok I'm a geek I'll go back to work now...

LICK MY DISEASED ANAL CUNT WIPE, FAGGOT!

[Anonymous Coward]

U = TEH FAGG0RTZ!!!1

Re:what's the point?

[edxwelch]

You actually have just proved the origional posters point: "64-bit is not faster than 32-bit in and of itself", and you say that your app is faster because of the extra registers. Extra registers are not related to 64-bitness... some one could just as easily bring out a 32-bit processor with extra registers and it would equally run faster ;)

Re:what's the point?

[Juergen Kreileder]

Good point :-)

AMD64 3200+ ? : install Mandrake 9.2RC1 amd64

[stock]

A hot tip for AMD64 3200+ and AMD64 3400+ owners. Install Mandrake 9.2RC1 amd64 and here's my results at mandrake's :

http://www.mandrakeclub.com/modules.php?op=modload &name=Splatt_Forum&file=viewtopic&topic=16806&foru m=9 some CPU info :

# cat /proc/cpuinfo
processor : 0
vendor_id : AuthenticAMD
cpu family : 15
model : 4
model name : AMD Athlon(tm) 64 Processor 3200+
stepping : 8
cpu MHz : 2199.873
cache size : 1024 KB
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext lm 3dnowext 3dnow
bogomips : 4390.91
TLB size : 1088 4K pages
clflush size : 64
address sizes : 40 bits physical, 48 bits virtual
power management: ts fid vid ttp
#
# openssl speed rsa
OpenSSL 0.9.7b 10 Apr 2003
built on: Tue Sep 9 14:19:06 CEST 2003
options:bn(64,64) md2(int) rc4(ptr,char) des(idx,cisc,16,int) aes(partial) idea(int) blowfish(ptr2)
compiler: gcc -fPIC -DOPENSSL_THREADS -D_REENTRANT -DDSO_DLFCN -DHAVE_DLFCN_H -DOPENSSL_NO_KRB5 -DOPENSSL_NO_ASM -DL_ENDIAN -DTERMIO -O2 -pipe -Wall -Wall -DMD32_REG_T=int
available timing options: TIMES TIMEB HZ=100 [sysconf value]
timing function used: times
sign verify sign/s verify/s
rsa 512 bits 0.0004s 0.0000s 2588.5 28043.6
rsa 1024 bits 0.0018s 0.0001s 543.8 9775.9
rsa 2048 bits 0.0112s 0.0003s 89.0 2952.3
rsa 4096 bits 0.0761s 0.0012s 13.1 845.5
#

can anyone beat those 28043.6 512bit rsa verify/sec??

Robert