OT: Cinebench on opteron and G5

[squidinc]

have you seen the price of those things?? damn... :D

interesting test though, not much in it really, I guess with the AMD system you'd get more for your money... but I'll leave it at that...

( don't flame me damnit :D )

[Erik Heyninck]

The more they try to make the best and the fastest, the better for all users!

Better so than Steve Jobstone and Bill Gatesrock playing with mechanical dinosaurs in Siliconstone valley!

Long live art, whether it's made with a G3,4 or 5, or a Celeron, PIII, PIV, or whatever AMD-based machine.

May those of you who have such a machine be happy with it, and make lots of exciting new things with it.

[LucentDreams]

good to see being mainly a C4D guy it owuldn't matter which I had, the results that catch my eye though are the PS ones considering how closely aplle and adobe worked to optimize, mayb CS is a different story

[Srek]

Hi Bart,
sadly they don't use the CINEBENCH the way it is intended to use, so a real comparison to other available results is not possible.
I happen to have a Dual Opteron 2 GHz in front of me now and it is rendering away at a blistering CB of 531 making him slightly faster then the Dual G5 2 GHz according to Mashs results
http://www.imashination.com/bench.html

Cheers

Srek

[brammelo]

Hi Bjorn,

Thanks for the info. However, both results match up pretty good - 524 for the G5 and 531 for the Opteron, that's a difference of around 1,5%. The world used to look a whole lot different for Mac owners :)

In what way did the guys over at bare feats use Cinebench improperly?

Cheers,
BaRa

[LucentDreams]

Originally posted by Srek
Hi Bart,
sadly they don't use the CINEBENCH the way it is intended to use, so a real comparison to other available results is not possible.
I happen to have a Dual Opteron 2 GHz in front of me now and it is rendering away at a blistering CB of 531 making him slightly faster then the Dual G5 2 GHz according to Mashs results
http://www.imashination.com/bench.html

Cheers

Srek

man the difference in results is soo minimal at that stage ;) especially when you consider my sytem can't make 150, so uhm 4 times as much, I don't care which I use for rendering I want one of something, I'm just witing to see who will give me a free system first ;) uhm yeah thats it.

[Srek]

Originally posted by brammelo
Hi Bjorn,

Thanks for the info. However, both results match up pretty good - 524 for the G5 and 531 for the Opteron, that's a difference of around 1,5%. The world used to look a whole lot different for Mac owners :)

In what way did the guys over at bare feats use Cinebench improperly?

Cheers,
BaRa
They are only comparing the time in seconds instead of using the Cinebench value. Since most people post the CB values you can't compare them to results in seconds. For Open GL the time can be completely misleading, since on fast machines some tests will run more then one time.
I know the difference isn't real worth mentioning, nevertheless i like to be correct in what i say ;)

Cheers

Srek

[Thalaxis]

Originally posted by Srek
Hi Bart,
sadly they don't use the CINEBENCH the way it is intended to use, so a real comparison to other available results is not possible.
I happen to have a Dual Opteron 2 GHz in front of me now and it is rendering away at a blistering CB of 531 making him slightly faster then the Dual G5 2 GHz according to Mashs results


Have you guys been able to make any decisions on AMD64
support yet, or are you still crippled in that regard by tool
availability, or more precisely the lack thereof?

I'm quite curious as to whether or not the AMD64 optimizing
compilers are any good, and also whether or not AMD's engineers
were talking out of their asses when they described the benefits
of the extra registers... :)

[AdamT]

Hi Bjorn,

When you compare the Opteron and G5 scores you're obviously using the G5 optimized CB. What about optimizations for the Opteron?

In either case a dual (or quad :)) Xeon system is still faster than either a G5 or Opteron system. Anyway I agree with Kai--it's great that there are now fast systems available for both platforms.

[Srek]

We still have to wait a bit for optimisation on Opteron. Like with G5 the current tools are not realy up to our standards yet. I for one will gladly miss some percent of processor power in exchange for keeping CINEMA as stable as possible.
Cheers
Srek

[Thalaxis]

That's about what I expected. Thanks for the reply!

[AdamT]

Hmm, I wonder if we can expect the same sort of gains with optimized Opteron code as you were able to achieve with the G5, i.e., around 20-30% speedup?

[Brent Turbo]

Coming from a completely impartial user of both PCs and Macs (me):

Barefeats can be classified as a "Mac Evangelist," type site, and their Mac benchmarks have, in the past, been far out of step (as in, waaay higher) with the benchmarks that other people were getting with the same systems.

On the topic of benchmarks--work smart, do creative stuff, and stop worrying about whether your computer is the fastest one on the market. Benchmarks are just computer wang-measuring contests, and as we all know, it's not the length, it's how you swing it.

[Thalaxis]

Originally posted by AdamT
Hmm, I wonder if we can expect the same sort of gains with optimized Opteron code as you were able to achieve with the G5, i.e., around 20-30% speedup?

That's what the AMD folks have stated.

[Srek]

Originally posted by AdamT
Hmm, I wonder if we can expect the same sort of gains with optimized Opteron code as you were able to achieve with the G5, i.e., around 20-30% speedup?

This depends heavily on what operations the Opteron can do faster or better and how they can be used inside of CINEMA.
An over all speedup of 20-30% seems unrealisticaly high to me.

Cheers
Srek

[Thalaxis]

That's precisely why I asked. The fact that AMD's saying it doesn't
make it true :)

I guess we'll see, eh?

[AdamT]

Originally posted by Brent Turbo
On the topic of benchmarks--work smart, do creative stuff, and stop worrying about whether your computer is the fastest one on the market. Benchmarks are just computer wang-measuring contests, and as we all know, it's not the length, it's how you swing it.
True, but not too helpful if you're in the market for a new computer and you're looking for the best bang/buck.

[CGTalk Moderation]

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.