Here it is: LW and Athlon 64 benchmarks!

[SLAYER]

Hell yeah!

http://www.tomshardware.com/cpu/20030923/athlon_64-45.html

[evenflcw]

People keep saying Toms Hardware is unreliable, this time I'm about to agree.

Just a couple of month ago Toms Hardware themselves showed how the 200FSB P4s really gave the XPs a good wipping, a P4 2.40Ghz 200FSB easily beating a Athlon XP 3200+! In this new benchmarks the numbers seem reversed somehow in some tests. Just compare the Cinema4D8.1 benchmarks from this old article (may 21) (http://www.tomshardware.com/cpu/20030521/800fsb-13.html) and the new one(sept 23) (http://www.tomshardware.com/cpu/20030923/athlon_64-46.html).

The LW benchmarks don't state what scene was use, but if was the same in both tests, then the same discrepancies exist. Test for seti@home and 3dsmax benchmark however match up over time pointing that some tests weren't done right this time or last time around.

[Freak]

I'd like to see LW8 tested, perhaps with a Linux node on the 64Bit
clients.. :)

I'm guessing Chipsets are still not quite where they should be,
usually takes 2-3 generations to get it together......

It's a shame that the P4's still beat the New 64bit machines...
(be it in 32Bit code) for AMD to really do damage to Intel, they need to beat Intel hands down.... Clock speeds do count when it comes to rendering, in fact it's about the only thing that counts.

BTW it does mention the render settings and scene used...

Considering that the P4's will be quite a lot cheaper, How many people are going to move over to a 64bit platform, which as yet does not really exist. (Software, OS) Until it does sales will be quite slow. It is a shame that AMD had to keep slower clock speeds.. Because to large Government departments,
Intel and Clockspeeds are about the only language they speak.

Trying to convince them to buy a slower more incompatible,
New technology, is like trying to sell them ****...

VT3 and LW8 will fair better with the AMD64, but still.....
I think AMD have bet the farm on this one, and i'm not convinced
they will be farming for that much longer. :(
Which means Intel will go on to dominate the world as usual.
(#^#@!@^#%^@%!!#@#) !!!!!

What we need is for the troubled Motorola, to join up with IBM and AMD to create a real force....

Still i'll move to AMD64bit as soon as i'm happy with the Mobo's and chipsets. Because i'd always support the underdog, rather than the IT Mafia....

[Exper]

evenflcw... totally agree...
Toms Hardware is an anty-AMD site (no matter if the number speak clear): overclocked P4 against plain Athlon64 on 32bit Windows!!!

Read here:
AMD's Athlon 64 processor - Lightwave rendering
http://www.tech-report.com/reviews/2003q3/athlon64/index.x?pg=13

AMD's Athlon 64 processor
http://www.tech-report.com/reviews/2003q3/athlon64/index.x?pg=1

and here:
AMD Athlon 64 & Athlon 64 FX - It's Judgement Day - 3D Rendering
http://www.anandtech.com/cpu/showdoc.html?i=1884&p=16

AMD Athlon 64 & Athlon 64 FX - It's Judgement Day
http://www.anandtech.com/cpu/showdoc.html?i=1884

Prices:
Athlon 64 3200+ will sell for $417 in 1,000 unit quantities
Athlon 64 FX 51 will be priced at $733 in 1,000 unit quantities
Pentium 4 3.2 EE will sell for around $740 in 1,000 unit quantities

Bye

[tallscot]

Any dual Opteron benchmarks with Lightwave yet?

These numbers I see for the Athlon 64 are slow compared to the dual Xeon and dual G5.

[CaptainKirk]

You can't expect a single Athlon chip to be faster that two xeons or even 2 crappy G5s. But it obviously will beat a single xeon and could probably beat a single G5 while rendering a Maya and Lightwave scene simultaneously.

Dual Opteron Benchmarks have been out for months and it's a little slower than dual xeon. The newest 2.2 Ghz Opterons could be same or a little faster even.

It does beat xeons at photoshop and Maya ( by a wide margin too )

[WizCraker]

Current benchmarks are useless now as the testing are with an application that is 32bit, you will only get the same or a small % increase when testing processors that are 32-64bit against just pure 32bit. If the application in question was not written for 64bit instructions, there might be a slight improvement in some scenes in the rendering, but until MS has shipped their 64bit XP for AMD out to general public [A firm or testing lab utilizing a beta or Release Canidate can not reflect actual results that can be achived in a Locked Release, Final optimization always occurs and last minute changes happen when a relase goes from a RC to Gold] and Newtek recompiles Lightwave for the x86-64bit architecture I would suggest you don't trust these benchmarks until at least two of the three are running in 64bit, that is Hardware [of course you need this one] and the OS. As for the Apple G5, Apple's website clearly states the G5 ships on their 32bit build of OS X, and that the 10.3 release due out later this year will have full optimization for the G5 as stated on their developer site.

So again wait for the releases then retest to a result that is running in actual 64bit.

[tallscot]

You can't expect a single Athlon chip to be faster that two xeons or even 2 crappy G5s. But it obviously will beat a single xeon and could probably beat a single G5 while rendering a Maya and Lightwave scene simultaneously.

In other words, it's useless.

Single Xeon? What's the point of that, even if I could get one?

LW Raytrace:
Dual 3.06 Ghz Xeon - 59 seconds
Dual 2 Ghz G5 - 63 seconds (not optimized LW and not Panther)

I don't think that G5 is crappy at all.

[Tronam]

Tom's Hardware is frequently flaky and unreliable. They will often rush out their articles as quickly as possible to try and beat everyone else to publication, but there are always mistakes all over the place that have to get patched up over time. Explain this one to me:

Here are the benchmarks for Cinema 4D doing a "dragon" scene (why he doesn't use Cinebench like everyone else is beyond me)...

http://www6.tomshardware.com/cpu/20030923/athlon_64-46.html


His summary then reads "The P4's HyperThreading really struts its stuff with this benchmark: the performance of the Athlon 64 FX-51 trails by 19.6 percent."

Here is the benchmark image:

http://www.tronam.com/images/newtek/chart_cinema.gif


Now, I may be overwhelmingly dense, but what in the hell is he talking about? :confused:

-Tronam

[prospector]

Now, I may be overwhelmingly dense,


It's people like you that give Toms a bad name.....
Don't you know that you should never check the printout graphic...
you should only read what he says.
:)

don't get so into the numbers.:D

[Lightwolf]

Originally posted by tallscot
Any dual Opteron benchmarks with Lightwave yet?

Check www.blanos.com, unfortunately, last time I checked, there weren't many LW 8 benchmarks yet (yup, the dual Xeon is mine ;) ).

Cheers,
Mike