Dual 2Ghz G5 vs Dual 2Ghz Opteron.....spin?

[Exposed]

This a link to a G5/Opteron shootout:
http://www.barefeats.com/g5op.html

Two of the four graphs were of the same result, shown differently. The author admits the 64-bit architecture of the Opteron was not taken advantage of, whereas the multimedia benches run were G5 optimized.

Similar benches run previously with inclusion of regular Athlon 2.4Ghz, Pentium 4 3.0Ghz, and Xeon 2.4 Ghz:
http://www.barefeats.com/pentium4.html

I suspect the Opteron benches were deliberately shown in their worst light, rather than their best (64 bit optimized applications).

 

[RussSchultz]

Wouldn't be the first time for Apple to stack their benchmarks.

 

[see colon]

"You'll notice Photoshop results are missing. My bad. I forgot to bring my PS 7 installer CD on my pilgrimage. I hope to add those results in a few days. I predict the G5 will win those two tests."
------------------------------------------------------------------------------------
ps7? shouldn't they be testing with the newest version?


"Wouldn't be the first time for Apple to stack their benchmarks."
------------------------------------------------------------------------------------
this isn't official apple benchmarks. the official benches should be out soon, though. i know this because... umm... if i tell you i'll get fired. but it's funny ;P
c:

 

[Entropy]

This a link to a G5/Opteron shootout:
http://www.barefeats.com/g5op.html

Two of the four graphs were of the same result, shown differently. The author admits the 64-bit architecture of the Opteron was not taken advantage of, whereas the multimedia benches run were G5 optimized.

Similar benches run previously with inclusion of regular Athlon 2.4Ghz, Pentium 4 3.0Ghz, and Xeon 2.4 Ghz:
http://www.barefeats.com/pentium4.html

I suspect the Opteron benches were deliberately shown in their worst light, rather than their best (64 bit optimized applications).

Not likely.
This guy is quite naive in his benchmarking, but he is not particularly biased. He simply shows the results he gets with very little analysis. If you snoop around on his site, you'll find a number of instances where Apple machines get creamed in games benchmarks. He has no qualms about showing such results. He is old enough to be a grandfather, and he doesn't tend to be very judgemental about the results he gets. He seems to be more curious than anything. Not banging drums.

The G5 really is a very fast machine for a large number of problems.
Once IBMs xlf and xlc sees more widespread use, (or more importantly, when gcc gets smarter about 970 instruction grouping and other peculiarities) it will get faster still.

That's not to say that the Opterons are bad machines or for that matter that there aren't codes where the P4s excel. It really shouldn't be a surprise that different architectures show strengths in different areas. What was a surprise was how good the G5 northbridge turned out to be.