About the “G4, G5, and AMD Shootout and Speed Testing in General…
I want to thank the folks who have dropped by the site to take a gander at the “G4, G5, and AMD Shootout” in The ComputerZone. And a special thanks to Mike at the “Accelerate your Mac” website. Mike really has a hell of a good website and does us Mac owners a huge public service by running it. My G4 iMac and my MDD Power Mac today are upgraded because of the material at Mike’s site. It’s really nice to have the option to upgrade these machines; without Mike and people like him, that option might not exist.
Sometimes, you just get lucky. I happened to be in the right place at the right time to perform a performance comparison between two older but recent generation Power Macs, a relatively current AMD powered XP machine, and one of the new G5 computers from Apple. (This type of thing is something I’ve always wanted to do.) The comparison was especially germane because the G5 and the AMD chip ran at roughly the same clock speed and video cards were matched up at all the machines at my end. True, it would have been a little better test of relative CPU strength had the G5 had a Radeon 9000 video card; but there was nothing I could do to change that since I didn’t have a G5 in the house. The G5 benchmarks came from chaosmint.com. Thanks to everyone involved in that site for posting them.
I’d like to say a few words here about the tests. Hopefully, you’ve seen some of the comments at xlr8yourmac.com about running Cinebench on the G5. Obviously, the code is not optimized for the machine. Folks involved with different aspects of the benchmark believe that once they optimize it for the G5, the numbers will go much higher. That’s good and great for bragging rights. But what does it mean to me today if I buy the G5? That’s the question I was after. Since my primary use for these machines is video and graphics, I felt that a rendering test like Cinebench was especially applicable. I’ve been hearing that Cinebench runs better on PC’s. Is that because the code is optimized better for PC’s or are they inherently faster at rendering than Mac’s are? Certainly, you would think that the Cinebench code would be optimized for the G4 by now; and as I sat and watched the test run on both sets of machines, it’s hard to argue that for the most part the AMD is not significantly faster. Does that mean I’m going to go back to AMD/Intel/Windows for my basic computer needs? Hell NO! Decades of running computers, working graphics and video on them, putting them together, and troubleshooting myself every single problem that occurred with them all convinced me that while speed was desirable, productivity is not so narrowly defined. Macs create a more desirable, integrated working environment than anything on the PC side, and that’s why I’ve spent huge amounts of money over the last couple of years converting over to them. I knew Macs were slower on some counts. But a slower machine that works—that you even have fun at—is worth more in my book than a faster machine I’m cussing at because I’ve got to troubleshoot it once more.
As I said in the test article, while there are some places where the AMD chip outperformed the 1.6 Ghz G5, they appear to be fairly evenly matched. (This would put the 1.6 Ghz G5 even today on par and perhaps slightly ahead of a 2Ghz P4.) Optimized Cinebench code will probably close the gap up a bit further and perhaps push the G5 ahead. But the real story won’t be told until applications like Final Cut Pro, DVD Studio Pro, Photoshop, In Design, Quark, and perhaps Mac games become truly optimized for the G5 running on a 64 bit operating system. That’s some years down the pike. This story is just beginning.
As for the story on this website, I’ll add dual 1.42 Ghz G4, 1.8 Ghz G5, and dual 2 Ghz G5 data to the plots as soon as I can find them or sweet-talk someone into letting me run Cinebench on their machine. Check back every now and then to see what I’ve found and feel free to send me any references.
Sometimes, you just get lucky. I happened to be in the right place at the right time to perform a performance comparison between two older but recent generation Power Macs, a relatively current AMD powered XP machine, and one of the new G5 computers from Apple. (This type of thing is something I’ve always wanted to do.) The comparison was especially germane because the G5 and the AMD chip ran at roughly the same clock speed and video cards were matched up at all the machines at my end. True, it would have been a little better test of relative CPU strength had the G5 had a Radeon 9000 video card; but there was nothing I could do to change that since I didn’t have a G5 in the house. The G5 benchmarks came from chaosmint.com. Thanks to everyone involved in that site for posting them.
I’d like to say a few words here about the tests. Hopefully, you’ve seen some of the comments at xlr8yourmac.com about running Cinebench on the G5. Obviously, the code is not optimized for the machine. Folks involved with different aspects of the benchmark believe that once they optimize it for the G5, the numbers will go much higher. That’s good and great for bragging rights. But what does it mean to me today if I buy the G5? That’s the question I was after. Since my primary use for these machines is video and graphics, I felt that a rendering test like Cinebench was especially applicable. I’ve been hearing that Cinebench runs better on PC’s. Is that because the code is optimized better for PC’s or are they inherently faster at rendering than Mac’s are? Certainly, you would think that the Cinebench code would be optimized for the G4 by now; and as I sat and watched the test run on both sets of machines, it’s hard to argue that for the most part the AMD is not significantly faster. Does that mean I’m going to go back to AMD/Intel/Windows for my basic computer needs? Hell NO! Decades of running computers, working graphics and video on them, putting them together, and troubleshooting myself every single problem that occurred with them all convinced me that while speed was desirable, productivity is not so narrowly defined. Macs create a more desirable, integrated working environment than anything on the PC side, and that’s why I’ve spent huge amounts of money over the last couple of years converting over to them. I knew Macs were slower on some counts. But a slower machine that works—that you even have fun at—is worth more in my book than a faster machine I’m cussing at because I’ve got to troubleshoot it once more.
As I said in the test article, while there are some places where the AMD chip outperformed the 1.6 Ghz G5, they appear to be fairly evenly matched. (This would put the 1.6 Ghz G5 even today on par and perhaps slightly ahead of a 2Ghz P4.) Optimized Cinebench code will probably close the gap up a bit further and perhaps push the G5 ahead. But the real story won’t be told until applications like Final Cut Pro, DVD Studio Pro, Photoshop, In Design, Quark, and perhaps Mac games become truly optimized for the G5 running on a 64 bit operating system. That’s some years down the pike. This story is just beginning.
As for the story on this website, I’ll add dual 1.42 Ghz G4, 1.8 Ghz G5, and dual 2 Ghz G5 data to the plots as soon as I can find them or sweet-talk someone into letting me run Cinebench on their machine. Check back every now and then to see what I’ve found and feel free to send me any references.
posted by Andy Foster at
3:20 AM
0 Comments:
Post a Comment
<< Home