Athlon 64 vs. Apple G5 Systems -- Athlon64 cleans house

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

dullard

Elite Member
May 21, 2001
26,185
4,844
126
Originally posted by: Eug

Anyways, I'm not surprised at all that an Opteron 246 is faster than a G5 2.0. I've been saying all along that a G5 2.0 is probably more in the speed range of an Opteron 244.

Other tests:
After Effects: The G5 cleans house usually, although x86 wins sometimes too.
Cinema4D: The Xeon with HT cleans house. (I haven't seen Opteron tests yet)
UT2003: The Athlon 64 cleans house.
ARC2D: The G5 cleans house (although I haven't seen the fastest Opterons yet running it), if you use IBM's compilers (vs. Intel's, etc).
BLAST: The G5 cleans house.
RenderMan: The 3 GHz dual Xeon is about as fast as the dual G5.
I agree with you Eug, but it is time now to consider newer Xeon processors. There is a 3.2 GHz Xeon that also has double the cache that the 3.06 GHz Xeon had in those tests you mention. Dell has been selling them for a while now, I don't know if you can get it anywhere else. Unfortunately no site EVER reviews the new Xeons right after their release - meaning we have to estimate. A good estimate would be to add a 4% boost for the faster clock speed (slightly less than the theoretical gain of 4.6%, 3.2/3.06 = 1.046). The extra cache is much harder to estimate though. The performance change ranges from a few percent worse to 31% better in SETI. You mentioned a few programs, I'll try my best to find benches with the new cache. Cinema 4D - 0% change, UT2003 - 11.7% faster, Photoshop - 6.9% faster.

Now combine the two effects: higher frequency and double cache. Lets use photoshop as an example. 1.04 * 1.069 = 1.112. This new Xeon processor should be 11% faster. Using your link the 3.2 GHz Xeon scores roughly 490*1.11 = 544 and the G5 scores 555. It really is a tie (considering I'm estimating things and being conservative the 3.2 GHz Xeon could be slightly higher than the G5). A tie doesn't result in a "read it and weap" post.

Of course the instant Apple has a faster G5, I'd hope you also update your benchmarks as soon as possible.
 

Eug

Lifer
Mar 11, 2000
24,167
1,812
126
Originally posted by: dullard
Originally posted by: Eug

Anyways, I'm not surprised at all that an Opteron 246 is faster than a G5 2.0. I've been saying all along that a G5 2.0 is probably more in the speed range of an Opteron 244.

Other tests:
After Effects: The G5 cleans house usually, although x86 wins sometimes too.
Cinema4D: The Xeon with HT cleans house. (I haven't seen Opteron tests yet)
UT2003: The Athlon 64 cleans house.
ARC2D: The G5 cleans house (although I haven't seen the fastest Opterons yet running it), if you use IBM's compilers (vs. Intel's, etc).
BLAST: The G5 cleans house.
RenderMan: The 3 GHz dual Xeon is about as fast as the dual G5.
I agree with you Eug, but it is time now to consider newer Xeon processors. There is a 3.2 GHz Xeon that also has double the cache that the 3.06 GHz Xeon had in those tests you mention. Dell has been selling them for a while now, I don't know if you can get it anywhere else. Unfortunately no site EVER reviews the new Xeons right after their release - meaning we have to estimate. A good estimate would be to add a 4% boost for the faster clock speed (slightly less than the theoretical gain of 4.6%, 3.2/3.06 = 1.046). The extra cache is much harder to estimate though. The performance change ranges from a few percent worse to 31% better in SETI. You mentioned a few programs, I'll try my best to find benches with the new cache. Cinema 4D - 0% change, UT2003 - 11.7% faster, Photoshop - 6.9% faster.

Now combine the two effects: higher frequency and double cache. Lets use photoshop as an example. 1.04 * 1.069 = 1.112. This new Xeon processor should be 11% faster. Using your link the 3.2 GHz Xeon scores roughly 490*1.11 = 544 and the G5 scores 555. It really is a tie (considering I'm estimating things and being conservative the 3.2 GHz Xeon could be slightly higher than the G5). A tie doesn't result in a "read it and weap" post.

Of course the instant Apple has a faster G5, I'd hope you also update your benchmarks as soon as possible.
Yep, I agree. The problem is finding somebody with a dual Opteron 246 or dual Xeon 3.2. Most of the stuff I've posted is a bunch of guys benchmarking their own systems (home or work) using specific guidelines, with numbers confirmed by others around the net.

There are lots of home users with dual G5s out there. (Including just moderate Photoshop users, etc., even though I tell them the cheaper G5 1.8 should be fine if they want a Mac. Why they all get dual G5s I don't quite understand, but it's true.) However, finding people with dual Xeons is a lot harder. And finding people with the latest dual Xeon 3.2 is next to impossible.

So I just report what numbers are available. You're right, with the new faster CPUs, more tests would be won for x86 (except for BLAST, since the G5 absolutely destroys x86 systems, likely because of Altivec optimizations).
 

Go3iverson

Senior member
Apr 16, 2000
273
0
0
Hi all,

It's always rough trying to find a fair comparison between a Mac and a PC. I must say, not to start a war, but this set of results from PCWorld really seems off to me. By simply looking over what apps they used as "fair" game between the two platforms I already had objections.

Word can be slower on a Mac than a PC, but I've never had anything take that long. It seems logical that a Microsoft product would run quicker on x86 architecture.

Adobe doesn't make Primiere for Mac any longer. Reason being, no one on the Mac side used it. I'm sure it had its fans, but I'm sure Windows 95 has its fans as well. Final Cut Pro is faster and considerably better, especially for the cost. Now I never used Primiere, but people that I know who have on Macs always said that FCP was a far superior product and simply that Primiere was slow.

The Photoshop test was pretty even. I'll never notice that 1 second difference from 1st to 2nd place. ;)
Also, I didn't have time to go through the whole article, but was the Photoshop G5 plug-in used? I know it says version 7.0.1, but I thought Adobe had a update prior to this.

No matter what, these are two beasts of computers, the Athlon 64 and the G5. It'll be nice to see AMD and Apple get a little more recognition for the products they produce.