PDA

View Full Version : Athlon 64


Personne
10-23-2004, 11:40 PM
If the price is not a factor which one between an Athlon 64 and a Pentium 4 is the best choice for 3D.

imashination
10-24-2004, 12:20 AM
Rendering - P4
3D view performance - A64

brudney
10-24-2004, 11:17 AM
i didn't know that a64 contributes to the viewport performance... :) how come?

and yeah, pentium seems to be better as far as rendering times are concerned. but again it largely depends on the app you use. ie 3ds max prefers p4 with it's high clock frequencies, whereas maya seems to like amd cpu architecture more... (but not necessarily a64; i rather speak of opteron vs. xeon in this case)

overall it's hard to say because render times are strongly influenced by the particular kind of scenes.

generally i guess i'd go for p4 (which doesn't mean a64 sucks; it's an awesome cpu, but somehow it manages better in games than in 3d)

imashination
10-24-2004, 12:46 PM
i didn't know that a64 contributes to the viewport performance... :) how come?


The A64 performs single threaded tasks very quickly, this means the 3d viewport part of an app. The HT bonus of the P4 allows it to pull ahead when rendering though.

brudney
10-24-2004, 01:02 PM
The A64 performs single threaded tasks very quickly, this means the 3d viewport part of an app. The HT bonus of the P4 allows it to pull ahead when rendering though.
yeah, theoretically. what about a real life situation? i actually doubt the difference in the viewport speed could be even noticed when comparing, say, a64 3400 and p4 3.2, given everything else (ram, video card...) is the same.

imashination
10-24-2004, 02:18 PM
yeah, theoretically. what about a real life situation? i actually doubt the difference in the viewport speed could be even noticed when comparing, say, a64 3400 and p4 3.2, given everything else (ram, video card...) is the same.

I don't do theoretical arguments, I leave that to other people. All of my comments are based on facts of what they actually do, not what they should do, or the manufacturers claim they do.

brudney
10-24-2004, 02:20 PM
I don't do theoretical arguments, I leave that to other people. All of my comments are based on facts of what they actually do, not what they should do, or the manufacturers claim they do.
heh, ok :)

arquebus
10-24-2004, 04:56 PM
If you dont mind me stretching this topic a bit, Im curious if the Mac G5 is used in rendering much. I saw a thing on tv how it was favored for use in super computer grids due to its processors ability to do floating point calculations.

Novakog
10-24-2004, 07:16 PM
Yeah, I have heard (not sure though) that Maya tends to favor AMD CPUs for rendering as well, but in all other cases for rendering Pentiums are better because of hyperthreading.

MadMax
10-24-2004, 07:23 PM
Rendering - P4
3D view performance - A64

Lightwave, AMD has top spot in rendering.

MAYA, AMD holds top spot in rendering.

Mental Ray seems to prefer AMD as well.


Now I do not personally use C4D or 3DS Max, however when the 2.4 ghz. AMD's came out, Anadtech's benchmarks had both C4D and Max split between AMD and Intel with each winning a portion of the benchmarks.

AMD's 2.6ghz is out, and Anandtech's review has the AMD competing extremely favorably.

In Max 6, AMD FX55 beat out the Intel 560 (3.6) or was a virtual tie on several of the tests and only slightly behind the EE version.

So I would say that to make the blanket statement that P4 is better for rendering isn't exactly accurate.

lots
10-24-2004, 08:11 PM
The problem with sites like Anandtech performing benchmarks with 3D apps, is they never look at all the big softwares out there at the same time. AND they never test the multitude of scenes that should be tested. Insted they'll test maybe two or three different scenes and call it a complete benchmark. That really bothers me...

The anandtech article you refer to is only evaluating 3DSMax. That doesnt speak for all the software out there...

Also holding the top spot in rendering is somewhat dependant on scene being rendered.. Dont you think? Making a "blanket statement" about AMD holding top spots in render just doesnt hold any water if its just one scene.

Dont get me wrong, I have used AMD since the first revision of the Athlon, and am planning on getting a dual opteron setup come Dec. I'm just pointing out that maybe the two are so close it doesnt matter so much which you get, other than personal pref, bragging rights, and the ability to prove to others you're set up is the best one.

MadMax
10-24-2004, 08:30 PM
Unfortunately "which is faster" is often determined as far as the forum types are concerned on specified benchmark scenes.

The Anandtech article I quoted listed 7 Max benchmarks, not just one, and the same with a previous article on C4D.

On both programs, AMD/Intel were split almost equally down the middle. Where AMD lost, it was by an amount so miniscule as to be unnoticed from the poit of view of someone actually working in a production environment. Likewise, where AMD won, the situation was the same. Margins to small to indicate without a benchmark to gauge fractions of a second.

The same situation holds true with the other fallacy, that Intel is better at media encoding. The situation is the same. Split dowen the middle with AMD and Intel being equally matched, the differences in win or loss less that a fracion of a second and not noticeable without a benchmark to tell you one was faster than the other.

From practical experience with Maya and Lightwave, I'd hands own take Opteron or A64.

Skipping all that and benchmarks, there are inumerable technological advantages to AMD over Intel whic is why their scores are so close, even though Intel has the "faster" chip.

Xeons with starved FSB's vs. Hypertransport, AMD runs cooler etc.

lots
10-24-2004, 08:52 PM
Oh I'm not sayin that the A64/Opteron are lacking in anyway (quite good designs, and with AMD's successful transition to 90nm its even better). I really like how they're designed, actually. Part of why im getting a pair :P. Just saying that I want these benchmarks to be more than what they are. Like I said, you hardly ever see any of the big apps compared head to head on the diff platforms. And if you do, they're only compared in a few areas. Plus in the article on anandtech, they had 5 benchmarks for 3dsmax ;) 7 if you count both versions of 3dsmax they used :P. But they still only tested 3DSMax, what about LW XSI Maya? It'd be nice to see how all that stacked up on the different platforms out there..

brudney
10-24-2004, 09:13 PM
But they still only tested 3DSMax, what about LW XSI Maya? It'd be nice to see how all that stacked up on the different platforms out there..yeah, that would be nice, but hey, it's anandtech! i actually don't expect them to make any serious 3d benchmarks. after all, it's a rather gaming oriented portal. i like them, their articles, benchmarks etc., it's a great source of knowledge, but don't expect to see lots of 3d benchmarks. in fact i'm happy that they actually do some!

face it, who besides a few CG freaks like us would like to scroll through dozens of meaning-god-knows-what tables showing some weird render times? if they actually did all those 3d benchmarks and i was a 16 year old guy who plays doom3 all the time, i would definitely avoid anandtech and find some more 'interesting' portal.

lots
10-24-2004, 09:38 PM
heh good point :)

Still doesnt mean that it wouldnt be handy to "us freaks" :)

CGTalk Moderation
01-19-2006, 05:00 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.