Inside a Quad G5

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

SLCentral

Diamond Member
Feb 13, 2003
3,542
0
71
Originally posted by: OdiN
Originally posted by: SLCentral
Originally posted by: OdiN
Originally posted by: SLCentral
Originally posted by: OdiN
Originally posted by: SLCentral
Originally posted by: MetalStorm
Great, I really can't wait for the benchmarks and for it to get it's ass handed to it by Opteron again. Super.

http://www.barefeats.com/mvdcpc.html

No Opteron tests yet, but the Quad completely owns all Intel and AMD (Athlon) offerings in the tests done.

You tool. They did 3 whole tests. And they are probably better suited to the Apple architecture. That and it takes Apple 4 cores to beat 2 AMD cores...and AMD's 2 cores are only 3 SECONDS behind the Apple's 4 cores in their beloved photoshop test...yeah...REALLY stomps on the AMD>

Er, if you look at Cinebench, the Quad did perform about 500 points "better" then the DC Athlon 2.2.

BTW, Barefeats has been considered one of the best, non-biased sources for comparing Apple to x86 systems.

And if you take 65% of the score in Cinebench and say that's roughly equivalent to a DC Apple...the DC AMD beats it.

They have shown their bias in that article.

If they wanted to test games, they could have easily put a 7800GT Mac Edition against a 7800GT on the PC's and run at say 1600x1200 and 1280x1024


Oh and yes the QC did beat the systems in video rendering....well big surprise it has 2 more cores to churn out frames. I could have told you the results of that test without even running it.

Thats exactly the point, the QC DID beat it. If you're a company needing the LEADING performance for an Apple platform, the QC is where to turn. We have yet to see Quad-core Opteron tests vs. a Apple QC system. I'm sure that the Opteron would stomp the Apple, but like I said, many companies need APPLE platforms.

EDIT: Notice how Barefeats tested the AMD/Intel solutions using both XP and XP-64? Real biased...

If a company "NEEDS" Apple...then why even bother comparing to a PC?

For people that dont NEED Apple, and are curious as to the comparison of performance between two systems?
 

OdiN

Banned
Mar 1, 2000
16,430
3
0
Originally posted by: SLCentral
Originally posted by: OdiN
Originally posted by: SLCentral
Originally posted by: OdiN
Originally posted by: SLCentral
Originally posted by: OdiN
Originally posted by: SLCentral
Originally posted by: MetalStorm
Great, I really can't wait for the benchmarks and for it to get it's ass handed to it by Opteron again. Super.

http://www.barefeats.com/mvdcpc.html

No Opteron tests yet, but the Quad completely owns all Intel and AMD (Athlon) offerings in the tests done.

You tool. They did 3 whole tests. And they are probably better suited to the Apple architecture. That and it takes Apple 4 cores to beat 2 AMD cores...and AMD's 2 cores are only 3 SECONDS behind the Apple's 4 cores in their beloved photoshop test...yeah...REALLY stomps on the AMD>

Er, if you look at Cinebench, the Quad did perform about 500 points "better" then the DC Athlon 2.2.

BTW, Barefeats has been considered one of the best, non-biased sources for comparing Apple to x86 systems.

And if you take 65% of the score in Cinebench and say that's roughly equivalent to a DC Apple...the DC AMD beats it.

They have shown their bias in that article.

If they wanted to test games, they could have easily put a 7800GT Mac Edition against a 7800GT on the PC's and run at say 1600x1200 and 1280x1024

Except that Apple has discontinued their 7800GT card :).

Regarding your dual-cord arguement, look at the scores. A Dual processor 2.7GHz Apple pretty much almost ties the Dual Core Athlon 2.2GHz, albeit losing a few points. Wheres the bias?

Except that they tested using a 7800GT in their benchmarks :confused:

Eek...my mistake. However, Barefeats is a private website that doesn't have every available resource. LEts face it, GPU's aren't cheap, and they can't have all of them. They're doing their best to accurately test, and we all know that an x86 processor will whip the G5 at gaming. It's a fact. When you take into account that most games only see one core, that makes it even worse for Apple. No one is arguing that.

Still their testing methodology is not the greatest.
 

SLCentral

Diamond Member
Feb 13, 2003
3,542
0
71
Originally posted by: OdiN
Originally posted by: SLCentral
Originally posted by: OdiN
Originally posted by: SLCentral
Originally posted by: OdiN
Originally posted by: SLCentral
Originally posted by: OdiN
Originally posted by: SLCentral
Originally posted by: MetalStorm
Great, I really can't wait for the benchmarks and for it to get it's ass handed to it by Opteron again. Super.

http://www.barefeats.com/mvdcpc.html

No Opteron tests yet, but the Quad completely owns all Intel and AMD (Athlon) offerings in the tests done.

You tool. They did 3 whole tests. And they are probably better suited to the Apple architecture. That and it takes Apple 4 cores to beat 2 AMD cores...and AMD's 2 cores are only 3 SECONDS behind the Apple's 4 cores in their beloved photoshop test...yeah...REALLY stomps on the AMD>

Er, if you look at Cinebench, the Quad did perform about 500 points "better" then the DC Athlon 2.2.

BTW, Barefeats has been considered one of the best, non-biased sources for comparing Apple to x86 systems.

And if you take 65% of the score in Cinebench and say that's roughly equivalent to a DC Apple...the DC AMD beats it.

They have shown their bias in that article.

If they wanted to test games, they could have easily put a 7800GT Mac Edition against a 7800GT on the PC's and run at say 1600x1200 and 1280x1024

Except that Apple has discontinued their 7800GT card :).

Regarding your dual-cord arguement, look at the scores. A Dual processor 2.7GHz Apple pretty much almost ties the Dual Core Athlon 2.2GHz, albeit losing a few points. Wheres the bias?

Except that they tested using a 7800GT in their benchmarks :confused:

Eek...my mistake. However, Barefeats is a private website that doesn't have every available resource. LEts face it, GPU's aren't cheap, and they can't have all of them. They're doing their best to accurately test, and we all know that an x86 processor will whip the G5 at gaming. It's a fact. When you take into account that most games only see one core, that makes it even worse for Apple. No one is arguing that.

Still their testing methodology is not the greatest.

Lets do the math. Obviously, a quad-core will not have 2x the performance of a dual-core. We all know that. But lets say take the Dual-Processor 2.5GHz and times the score by two (for Cinebench). That gives it 1314. Then take the DC 2.2Ghz Athlon in 64-bit mode, which gives it 1514. Clearly, the AMD processor has the lead. I'm not saying it doesn't. I'm saying that for many people, that slight advantage MAY not be as important as the operating system. OS X does have value.
 

OdiN

Banned
Mar 1, 2000
16,430
3
0
Meh...nobody has convinced me that OS X is any better than WinXP. I've seen Apple's OSes crash a lot more than any OS on my computers. Even Win98.
 

SLCentral

Diamond Member
Feb 13, 2003
3,542
0
71
Originally posted by: OdiN
Meh...nobody has convinced me that OS X is any better than WinXP. I've seen Apple's OSes crash a lot more than any OS on my computers. Even Win98.

Your opinion...
 

OdiN

Banned
Mar 1, 2000
16,430
3
0
Originally posted by: SLCentral
Originally posted by: OdiN
Meh...nobody has convinced me that OS X is any better than WinXP. I've seen Apple's OSes crash a lot more than any OS on my computers. Even Win98.

Your opinion...

No..I've pretty much seen this fact with my own eyes.
 

SLCentral

Diamond Member
Feb 13, 2003
3,542
0
71
Originally posted by: OdiN
Originally posted by: SLCentral
Originally posted by: OdiN
Meh...nobody has convinced me that OS X is any better than WinXP. I've seen Apple's OSes crash a lot more than any OS on my computers. Even Win98.

Your opinion...

No..I've pretty much seen this fact with my own eyes.

Probably a buggy system then.
 

OdiN

Banned
Mar 1, 2000
16,430
3
0
Originally posted by: SLCentral
Originally posted by: OdiN
Originally posted by: SLCentral
Originally posted by: OdiN
Meh...nobody has convinced me that OS X is any better than WinXP. I've seen Apple's OSes crash a lot more than any OS on my computers. Even Win98.

Your opinion...

No..I've pretty much seen this fact with my own eyes.

Probably a buggy system then.

Systems. I've worked with a bunch of them in graphics design departments at a couple of schools. I've been stuck using them in High School as well.
 

RaynorWolfcastle

Diamond Member
Feb 8, 2001
8,968
16
81
Originally posted by: Vegito
how is this thing.. my coworker wants to buy one to run matlab ...

AFAIK, matlab is single threaded so your coworker is better off with a fast single-core CPU (or dual core if he likes to multi-task a lot).
 

SLCentral

Diamond Member
Feb 13, 2003
3,542
0
71
Originally posted by: RaynorWolfcastle
Originally posted by: Vegito
how is this thing.. my coworker wants to buy one to run matlab ...

AFAIK, matlab is single threaded so your coworker is better off with a fast single-core CPU (or dual core if he likes to multi-task a lot).

If it's single-threaded, then you're absolutely right.
 

SLCentral

Diamond Member
Feb 13, 2003
3,542
0
71
Originally posted by: OdiN
Originally posted by: SLCentral
Originally posted by: OdiN
Originally posted by: SLCentral
Originally posted by: OdiN
Meh...nobody has convinced me that OS X is any better than WinXP. I've seen Apple's OSes crash a lot more than any OS on my computers. Even Win98.

Your opinion...

No..I've pretty much seen this fact with my own eyes.

Probably a buggy system then.

Systems. I've worked with a bunch of them in graphics design departments at a couple of schools. I've been stuck using them in High School as well.

Not sure what was going on then. I've owned about a dozen Macs in the past three years and none of them have had any problems, as well as any Mac that I've ever used.
 

thirdlegstump

Banned
Feb 12, 2001
8,713
0
0
Originally posted by: SLCentral
Originally posted by: OdiN
Originally posted by: SLCentral
Originally posted by: OdiN
Originally posted by: SLCentral
Originally posted by: OdiN
Meh...nobody has convinced me that OS X is any better than WinXP. I've seen Apple's OSes crash a lot more than any OS on my computers. Even Win98.

Your opinion...

No..I've pretty much seen this fact with my own eyes.

Probably a buggy system then.

Systems. I've worked with a bunch of them in graphics design departments at a couple of schools. I've been stuck using them in High School as well.

Not sure what was going on then. I've owned about a dozen Macs in the past three years and none of them have had any problems, as well as any Mac that I've ever used.

Ditto. I've owned too many and none ever exhibited problems of any kind. They always ran smooth and by the numbers.
 

jdogg707

Diamond Member
Jun 24, 2002
6,098
0
76
Macs do crash, but I promise you they crash a lot less than XP PC's. :)

Some people are know it all Mac haters who will not be convinced of anything. I personally would rather have an open mind and except that both systems have their advanatages and disadvantages. To this point, Macvs are better suited for some things while XP machines are better suited to others. I am looking forward to the switch to Intel chips as this will help bridge that difference.
 

MetalStorm

Member
Dec 22, 2004
148
0
0
Originally posted by: SLCentral
Er, if you look at Cinebench, the Quad did perform about 500 points "better" then the DC Athlon 2.2.

BTW, Barefeats has been considered one of the best, non-biased sources for comparing Apple to x86 systems.

Edit: If you look at Barefeats past test, you'll see that the PC's won most of the tests. Barefeats is NOT biased. I'm sure they will extend their testing to other platforms and tests as soon as they have the oppertunity to.

"We weren't able to match the clock speed exactly, but I'm giving you the results from those Dual-Core Macs and PCs made avaialable to us so far."

On top of that, the tests done are all programs that are widely used. AfterEffects is VERY much used, and the Quad beat all the other machines significantly.

Would you like to read my post and then reconsider how "unbiased" they really are?

Thank you.
 

SLCentral

Diamond Member
Feb 13, 2003
3,542
0
71
Originally posted by: MetalStorm
I put very little stock in those benchmarks, or more to the point the site that they are from, I've seen benchmarks from that site before and the results they have drawn have been far more than dubious.

While it's hard to say what settings or various sneeky things they have altered on the After Effects and Photoshop benchmarks to give the G5 the advantage it can easily be seen on their game benchmarks that they do not know their ass from their elbow - or rather then don't think you do. Notice how all the game benchmarks are done in a resolution of 1920x1080... Wait are these graphics card benchmarks? Oh, they're CPU benchmarks, the usual testing policy in that case is to run the game and take the GPU out of the equation by REDUCING the resolution... Hey, at least the G5 looks like it's almost as fast as the PCs there right?

Of course that's the idea, if they didn't make the GPU limit the FPS then the PCs, more to the point the Athlon 64s would be scoring around 120FPS in the DOOM3 benchmark, and wouldn't that make the G5 look bad!

So to conclude, if they can make such a hash of the game benchmarks, I really don't trust their other numbers. Not to mention benchmarking a QUAD rig against DUALs???

Sorry, I completely forgot to reply to this.

You cannot assume they altered things with AfterEffects and Photoshop and say they are biased. With that mindset, you can say any review is biased. Hell, according to that, all of AT's reviews are biased, becasue, well, theres a CHANCE they changed settings.

Games, I agree with you. Then again, everyone knows Macs aren't a gaming platform. They have tested in the past with multiple resolutions in games, but apparently they have not (yet) done so here.

Regarding quad vs. dual, Quad Macs are significantly more available then quad PC's are, as very few places SELL Quad PC's. Also keep in mind that this is their first testing of Quad's vs. PC's, and they said this is what they have SO FAR.
 

ShadowBlade

Diamond Member
Feb 11, 2005
4,263
0
0
Originally posted by: SLCentral
Originally posted by: OdiN
Originally posted by: SLCentral
Originally posted by: OdiN
Originally posted by: SLCentral
Originally posted by: OdiN
Originally posted by: SLCentral
Originally posted by: OdiN
Originally posted by: SLCentral
Originally posted by: MetalStorm
Great, I really can't wait for the benchmarks and for it to get it's ass handed to it by Opteron again. Super.

http://www.barefeats.com/mvdcpc.html

No Opteron tests yet, but the Quad completely owns all Intel and AMD (Athlon) offerings in the tests done.

You tool. They did 3 whole tests. And they are probably better suited to the Apple architecture. That and it takes Apple 4 cores to beat 2 AMD cores...and AMD's 2 cores are only 3 SECONDS behind the Apple's 4 cores in their beloved photoshop test...yeah...REALLY stomps on the AMD>

Er, if you look at Cinebench, the Quad did perform about 500 points "better" then the DC Athlon 2.2.

BTW, Barefeats has been considered one of the best, non-biased sources for comparing Apple to x86 systems.

And if you take 65% of the score in Cinebench and say that's roughly equivalent to a DC Apple...the DC AMD beats it.

They have shown their bias in that article.

If they wanted to test games, they could have easily put a 7800GT Mac Edition against a 7800GT on the PC's and run at say 1600x1200 and 1280x1024

Except that Apple has discontinued their 7800GT card :).

Regarding your dual-cord arguement, look at the scores. A Dual processor 2.7GHz Apple pretty much almost ties the Dual Core Athlon 2.2GHz, albeit losing a few points. Wheres the bias?

Except that they tested using a 7800GT in their benchmarks :confused:

Eek...my mistake. However, Barefeats is a private website that doesn't have every available resource. LEts face it, GPU's aren't cheap, and they can't have all of them. They're doing their best to accurately test, and we all know that an x86 processor will whip the G5 at gaming. It's a fact. When you take into account that most games only see one core, that makes it even worse for Apple. No one is arguing that.

Still their testing methodology is not the greatest.

Lets do the math. Obviously, a quad-core will not have 2x the performance of a dual-core. We all know that. But lets say take the Dual-Processor 2.5GHz and times the score by two (for Cinebench). That gives it 1314. Then take the DC 2.2Ghz Athlon in 64-bit mode, which gives it 1514. Clearly, the AMD processor has the lead. I'm not saying it doesn't. I'm saying that for many people, that slight advantage MAY not be as important as the operating system. OS X does have value.
[/quote]
I'll go ahead and continue the quote string
 

MetalStorm

Member
Dec 22, 2004
148
0
0
Originally posted by: SLCentral
Originally posted by: MetalStorm
I put very little stock in those benchmarks, or more to the point the site that they are from, I've seen benchmarks from that site before and the results they have drawn have been far more than dubious.

While it's hard to say what settings or various sneeky things they have altered on the After Effects and Photoshop benchmarks to give the G5 the advantage it can easily be seen on their game benchmarks that they do not know their ass from their elbow - or rather then don't think you do. Notice how all the game benchmarks are done in a resolution of 1920x1080... Wait are these graphics card benchmarks? Oh, they're CPU benchmarks, the usual testing policy in that case is to run the game and take the GPU out of the equation by REDUCING the resolution... Hey, at least the G5 looks like it's almost as fast as the PCs there right?

Of course that's the idea, if they didn't make the GPU limit the FPS then the PCs, more to the point the Athlon 64s would be scoring around 120FPS in the DOOM3 benchmark, and wouldn't that make the G5 look bad!

So to conclude, if they can make such a hash of the game benchmarks, I really don't trust their other numbers. Not to mention benchmarking a QUAD rig against DUALs???

Sorry, I completely forgot to reply to this.

You cannot assume they altered things with AfterEffects and Photoshop and say they are biased. With that mindset, you can say any review is biased. Hell, according to that, all of AT's reviews are biased, becasue, well, theres a CHANCE they changed settings.

Games, I agree with you. Then again, everyone knows Macs aren't a gaming platform. They have tested in the past with multiple resolutions in games, but apparently they have not (yet) done so here.

Regarding quad vs. dual, Quad Macs are significantly more available then quad PC's are, as very few places SELL Quad PC's. Also keep in mind that this is their first testing of Quad's vs. PC's, and they said this is what they have SO FAR.

The point I'm making is that they've tested the games in a biased manner, basically making the macs look not as bad as they should, and really dont provide the data to tell if the other benchmarks have been performed correctly or not.
I'm a scientist, when you test something or are trying to prove something you write down exactly how you tested and the various settings etc so that someone else can test under the same conditions and see if they get the same results and draw the same conclusions. That's how these things work. If data is not present it can be assumed to be default, however with the obvious choice of screen resolution in the game benchmarks, it is quite possible they are trying to hide something. It really does call their methodology in to question.

Think about it like this, with their photoshop and AE benchmarks, they don't tell you, or show you what image they are manipulating, you assume they are the same image of course but that's not stated, I would be surprised if they did have different images, so I'm not accusing them of cheating on that account. However you have no idea of the filters they used, for all you know they could have quite easily found a filter on the internet which has been heavily optimised for the G5 and used that, even if they were to say they tested 20 filters, they could still have cherry picked the filters to give the best results, and I do believe some of this has been going on.

Oh not to mention that I've seen benchmarks that completly reverse the tables, and they also come from sites that aren't quite so biased. Have a look on Anandtech, they did some quite interesting server benchmarks against the iserve or whatever it's called, quite ammusing.
 

SLCentral

Diamond Member
Feb 13, 2003
3,542
0
71
Originally posted by: MetalStorm
Originally posted by: SLCentral
Originally posted by: MetalStorm
I put very little stock in those benchmarks, or more to the point the site that they are from, I've seen benchmarks from that site before and the results they have drawn have been far more than dubious.

While it's hard to say what settings or various sneeky things they have altered on the After Effects and Photoshop benchmarks to give the G5 the advantage it can easily be seen on their game benchmarks that they do not know their ass from their elbow - or rather then don't think you do. Notice how all the game benchmarks are done in a resolution of 1920x1080... Wait are these graphics card benchmarks? Oh, they're CPU benchmarks, the usual testing policy in that case is to run the game and take the GPU out of the equation by REDUCING the resolution... Hey, at least the G5 looks like it's almost as fast as the PCs there right?

Of course that's the idea, if they didn't make the GPU limit the FPS then the PCs, more to the point the Athlon 64s would be scoring around 120FPS in the DOOM3 benchmark, and wouldn't that make the G5 look bad!

So to conclude, if they can make such a hash of the game benchmarks, I really don't trust their other numbers. Not to mention benchmarking a QUAD rig against DUALs???

Sorry, I completely forgot to reply to this.

You cannot assume they altered things with AfterEffects and Photoshop and say they are biased. With that mindset, you can say any review is biased. Hell, according to that, all of AT's reviews are biased, becasue, well, theres a CHANCE they changed settings.

Games, I agree with you. Then again, everyone knows Macs aren't a gaming platform. They have tested in the past with multiple resolutions in games, but apparently they have not (yet) done so here.

Regarding quad vs. dual, Quad Macs are significantly more available then quad PC's are, as very few places SELL Quad PC's. Also keep in mind that this is their first testing of Quad's vs. PC's, and they said this is what they have SO FAR.

The point I'm making is that they've tested the games in a biased manner, basically making the macs look not as bad as they should, and really dont provide the data to tell if the other benchmarks have been performed correctly or not.
I'm a scientist, when you test something or are trying to prove something you write down exactly how you tested and the various settings etc so that someone else can test under the same conditions and see if they get the same results and draw the same conclusions. That's how these things work. If data is not present it can be assumed to be default, however with the obvious choice of screen resolution in the game benchmarks, it is quite possible they are trying to hide something. It really does call their methodology in to question.

Think about it like this, with their photoshop and AE benchmarks, they don't tell you, or show you what image they are manipulating, you assume they are the same image of course but that's not stated, I would be surprised if they did have different images, so I'm not accusing them of cheating on that account. However you have no idea of the filters they used, for all you know they could have quite easily found a filter on the internet which has been heavily optimised for the G5 and used that, even if they were to say they tested 20 filters, they could still have cherry picked the filters to give the best results, and I do believe some of this has been going on.

Oh not to mention that I've seen benchmarks that completly reverse the tables, and they also come from sites that aren't quite so biased. Have a look on Anandtech, they did some quite interesting server benchmarks again the iserve or whatever it's called, quite ammusing.

I'm not sure about the testing methodology done by Barefeats, so theres no point in arguing about it. Regarding the benchmarks done by AT on the xSever, AT acknowledged the miserable performance was because OS X was not optimized for MySQL, not because of the actual hardware. That has nothing to do with a consumer enviornment, though.
 

MetalStorm

Member
Dec 22, 2004
148
0
0
SLCentral I really can not explain it any other way, so while you might not get it, I hope the others on here do, and hence take the benchmarks with a rather large grain of salt.
 

SLCentral

Diamond Member
Feb 13, 2003
3,542
0
71
Originally posted by: MetalStorm
SLCentral I really can not explain it any other way, so while you might not get it, I hope the others on here do, and hence take the benchmarks with a rather large grain of salt.

I completely understand what you are saying, I'm just saying you can't assume without confirming.

Here is a guide to his testing methodology that he used a couple years back. I doubt its the same, but it gives a good idea as to how he tests. I'm sure if you (or I) sent him an e-mail, he'd be glad to explain his testing methods.
 

imported_Lucifer

Diamond Member
Oct 12, 2004
5,139
1
0
Barefeats is not a biased site. Just take a look at a lot of the mac vs PC benchmarks. There are lots of tests that show the G5 getting it's ass handed to in some tests. A lot of the tests do show that one is better than the other though. The G5 is better at one thing, while the PC is better at the other.

I emailed Rob Art, barefeats site owner, a few months back regarding the Pentium D system he tested against the Dual processor G5. I got a reply within 30 minutes. Reason I emailed him is because he didn't have the full specs posted. He emailed me the specs, then corrected the web page by adding the full specs to it.