G5 Speed Tests
- Thread starter mindbend
- Start date
hulkaros
The Incredible...
Me too, thought at first that PC Magazine did a great job in this VS article until... I read here:
http://discuss.pcmag.com/n/mb/message.asp?webtag=pcmag&msg=28784.1&maxT=3
PC Mag after all is full of BS
Here are some for those that are too bored to read the whole discussion that goes over there:
"I thought the article was well-balanced overall, but comments about pricing were misleading.
It seems Mr. Dreier has selected BTO options in such a way as to make the prices as close as possible. Although Mr. Dreier claimed the G5 and Dell Precision 650 were "similarly configured (and priced)", the G5 can actually be had for much less than the Dell.
First, when I configured a Dell with options similar to those Mr. Dreier chose for the G5, the Dell priced out at $4,934 vs. $4,349 for the G5. These prices are not "similar." Second, the stock G5 is also much cheaper than a similarly configured Dell. Whereas the stock G5 costs $2,999, a base configuration of the Dell costs $3,855, and that's with a 32MB graphics card, 120GB hard drive, and 266MHz DDR RAM versus the 64MB graphics card, 160GB hard drive, and 400MHz DDR RAM equipped on the stock G5. Third, the decision to buy 2GB of RAM from Apple as a BTO option was something of a red herring and one most consumers would not make. RAM can be purchased much more cheaply elsewhere. Inclusion of this option makes the G5 seem much less price-competitive than it actually is.
The real fact is that, not only does the G5 outperform the Dell, but it does so for hundreds of dollars less."
Another one:
"2) "You can load the G5 with up to"... "new 160GB Serial ATA hard drives". That's misleading.. you can put dual 250 GB SATA drives in the G5, from Apple.
3) It was mentioned elsewhere that they messed up the count of fw/usb2 ports"
Some more:
"1. You give Avid Xpress's results in the pop-up, but not in the final numbers on the last page. Avid was much faster on the G5 and should be included on the final page of the review.
2. The Lightwave benchmarks you did were flawed. Lightwave requires that you manually set the threads above 1 if you have dual processors. Clearly, PCMag didn't do this because their results aren't close to what these systems get with those same benchmark scenes. The Lightwave results should be ignored."
And one last question which comes from something that PC Mag said in Photoshop part:
"Moreover, on the Windows system, loading the controls often took a minute or more. If these times are added back to the actual test times, both Macintosh computers would have clearly outperformed the Windows-based computer."
What are these Photoshop controls? And why one should not insert them into benchmarking times?
Not only G5 IS faster than a Dual XEON 3GHz but cheaper too!
I CANNOT WAIT FOR MY DUAL G5!
Apple rulez! Dull sux!
http://discuss.pcmag.com/n/mb/message.asp?webtag=pcmag&msg=28784.1&maxT=3
PC Mag after all is full of BS
Here are some for those that are too bored to read the whole discussion that goes over there:
"I thought the article was well-balanced overall, but comments about pricing were misleading.
It seems Mr. Dreier has selected BTO options in such a way as to make the prices as close as possible. Although Mr. Dreier claimed the G5 and Dell Precision 650 were "similarly configured (and priced)", the G5 can actually be had for much less than the Dell.
First, when I configured a Dell with options similar to those Mr. Dreier chose for the G5, the Dell priced out at $4,934 vs. $4,349 for the G5. These prices are not "similar." Second, the stock G5 is also much cheaper than a similarly configured Dell. Whereas the stock G5 costs $2,999, a base configuration of the Dell costs $3,855, and that's with a 32MB graphics card, 120GB hard drive, and 266MHz DDR RAM versus the 64MB graphics card, 160GB hard drive, and 400MHz DDR RAM equipped on the stock G5. Third, the decision to buy 2GB of RAM from Apple as a BTO option was something of a red herring and one most consumers would not make. RAM can be purchased much more cheaply elsewhere. Inclusion of this option makes the G5 seem much less price-competitive than it actually is.
The real fact is that, not only does the G5 outperform the Dell, but it does so for hundreds of dollars less."
Another one:
"2) "You can load the G5 with up to"... "new 160GB Serial ATA hard drives". That's misleading.. you can put dual 250 GB SATA drives in the G5, from Apple.
3) It was mentioned elsewhere that they messed up the count of fw/usb2 ports"
Some more:
"1. You give Avid Xpress's results in the pop-up, but not in the final numbers on the last page. Avid was much faster on the G5 and should be included on the final page of the review.
2. The Lightwave benchmarks you did were flawed. Lightwave requires that you manually set the threads above 1 if you have dual processors. Clearly, PCMag didn't do this because their results aren't close to what these systems get with those same benchmark scenes. The Lightwave results should be ignored."
And one last question which comes from something that PC Mag said in Photoshop part:
"Moreover, on the Windows system, loading the controls often took a minute or more. If these times are added back to the actual test times, both Macintosh computers would have clearly outperformed the Windows-based computer."
What are these Photoshop controls? And why one should not insert them into benchmarking times?
Not only G5 IS faster than a Dual XEON 3GHz but cheaper too!
I CANNOT WAIT FOR MY DUAL G5!
Apple rulez! Dull sux!
Even better!
The only thing I might add to defray some of that is that Lightwave is historically slow on the Mac for whatever reason. Especially in raytracing. The G5 will help, but LW needs a serious bit of reprogramming to run as well as it should on the Mac. This is prime opprtunity for competitors to pull ahead. Carrara is a nice mid-level program that could shine if it runs fast on the G5s.
I don't know what they are talking about with the Photoshop "controls". They can't mean the palettes or the filter option windows. How could you function if they took a minute or more to open? Does anyone know what the heck they are talking about?
The only thing I might add to defray some of that is that Lightwave is historically slow on the Mac for whatever reason. Especially in raytracing. The G5 will help, but LW needs a serious bit of reprogramming to run as well as it should on the Mac. This is prime opprtunity for competitors to pull ahead. Carrara is a nice mid-level program that could shine if it runs fast on the G5s.
I don't know what they are talking about with the Photoshop "controls". They can't mean the palettes or the filter option windows. How could you function if they took a minute or more to open? Does anyone know what the heck they are talking about?
hulkaros
The Incredible...
Nothing THAT important but a good read anyways...
http://www.macosxhints.com/g5/
I like the movie where he shows that in 10 secs he loads 10 apps (including heavy ones like iPhoto, iMovie, etc.)!!!
2-3 days left for my Dual G5! I cannot wait! My heart will fail from the excitement!!!
http://www.macosxhints.com/g5/
I like the movie where he shows that in 10 secs he loads 10 apps (including heavy ones like iPhoto, iMovie, etc.)!!!
2-3 days left for my Dual G5! I cannot wait!
My heart will fail from the excitement!!! Two Words: Holy and crap.
A real OS X machine is finally here. I have never considered G4s suitable for OS X (not trolling, just my opinion). The only reason I hung on the last three years is because I knew the hardware would finally catch up. The G5 is it. Once again, brilliant of Apple to anticipate the needs of their customers by risking it all and putting the cart (OS X) before the horse (G5). Nice job.
A real OS X machine is finally here. I have never considered G4s suitable for OS X (not trolling, just my opinion). The only reason I hung on the last three years is because I knew the hardware would finally catch up. The G5 is it. Once again, brilliant of Apple to anticipate the needs of their customers by risking it all and putting the cart (OS X) before the horse (G5). Nice job.
hulkaros
The Incredible...
Thanks arden but hurry up! My heart is going to explode from the excitement
And mindbend, I wish you the best in your going-to-buy-the-new-Dual G5
BTW did you see that video of 10 apps opening in 10 secs tops? WOW!
And this is with 10.2.8... Can you imagine the Dual G5 coupled with Panther in a few weeks? Crazy stuff!
And mindbend, I wish you the best in your going-to-buy-the-new-Dual G5
BTW did you see that video of 10 apps opening in 10 secs tops? WOW!
And this is with 10.2.8... Can you imagine the Dual G5 coupled with Panther in a few weeks?
Crazy stuff! Just a couple of observations. Not trying to be negative, but just noticing some things.
1. Those fast loading apps are mostly Apple's. This is not a criticism of the test, just an observation. It shows that proper programming can yield wicked results. However, Adobe Elements and Safari show no evidence of actually launching. They may very well be hidden in the background somwhere, but I could find no evidence that they actually launched (other than the dock indicator, which is enough I guess).
2. Elsewhere on his site, he has the launch time of Elements as 6 seconds, which is nice, very nice, but not as instant as the other "Applets". I suspect full blown Photoshop, InDesign, Illustrator will still take several seconds each. Currently, Photoshop takes 18 seconds to launch for me on a DP 1 gig, so if I can get down around six seconds, that's quite a speed bump. I can live with waiting for Photoshop for a few seconds. What has always been really annoying is waiting for the little application-wanna-bes to load in multiple bounces. I mean, waiting more than one second for System Prefs/iTunes/Terminal/Calculator to open is a slap in the face. Not anymore.
3. I did a quick, super-lame, app launch test similar to his movie. I didn't bother to do a screen movie cap, so you'll just have to go with my info. My apps are mail/isync/safari/explorer/dvd player/itunes/iphoto/ical/omnigraffle/fetch/termianl/sysprefs. They appeared to launch in fifteen seconds, but actually they were still loading even though the dock bounce effect had stopped. Actual launch times were closer to 25-30 seconds. With the apps cached and relaunched, it took no more than 25 seconds to fully launch them all. I suspect this is why we don't see Adobe Elements and Safari fully loaded in the osxhints movie. Still, the G5 should launch apps very fast, I'm sure. To be fair, my understanding is that Panther is a big help in this area as well, it's possible that a DP G4 could launch those sames apps almost as fast.
4. MP3 rip. Rob states that he can rip MP3s on the G5 at up to 16.4 speed. My G4 DP 1 gig can easily match that. It easily breaks 20x speeds and has peaked at 25x. Average is around 15-17x. I presume he was ripping right from the CD, which means the SuperDrive is a bottleneck, so I can't fault the G5 for that. I'd be curious to see how fast it would rip aiffs from the hard drive (not that anybody would pull aiffs from the CD and THEN rip them, but often times I create AAC files from AIFs created in CuBase).
These are about as harsh of criticisms as I can muster, which means the G5 appears to be a monster.
1. Those fast loading apps are mostly Apple's. This is not a criticism of the test, just an observation. It shows that proper programming can yield wicked results. However, Adobe Elements and Safari show no evidence of actually launching. They may very well be hidden in the background somwhere, but I could find no evidence that they actually launched (other than the dock indicator, which is enough I guess).
2. Elsewhere on his site, he has the launch time of Elements as 6 seconds, which is nice, very nice, but not as instant as the other "Applets". I suspect full blown Photoshop, InDesign, Illustrator will still take several seconds each. Currently, Photoshop takes 18 seconds to launch for me on a DP 1 gig, so if I can get down around six seconds, that's quite a speed bump. I can live with waiting for Photoshop for a few seconds. What has always been really annoying is waiting for the little application-wanna-bes to load in multiple bounces. I mean, waiting more than one second for System Prefs/iTunes/Terminal/Calculator to open is a slap in the face. Not anymore.
3. I did a quick, super-lame, app launch test similar to his movie. I didn't bother to do a screen movie cap, so you'll just have to go with my info. My apps are mail/isync/safari/explorer/dvd player/itunes/iphoto/ical/omnigraffle/fetch/termianl/sysprefs. They appeared to launch in fifteen seconds, but actually they were still loading even though the dock bounce effect had stopped. Actual launch times were closer to 25-30 seconds. With the apps cached and relaunched, it took no more than 25 seconds to fully launch them all. I suspect this is why we don't see Adobe Elements and Safari fully loaded in the osxhints movie. Still, the G5 should launch apps very fast, I'm sure. To be fair, my understanding is that Panther is a big help in this area as well, it's possible that a DP G4 could launch those sames apps almost as fast.
4. MP3 rip. Rob states that he can rip MP3s on the G5 at up to 16.4 speed. My G4 DP 1 gig can easily match that. It easily breaks 20x speeds and has peaked at 25x. Average is around 15-17x. I presume he was ripping right from the CD, which means the SuperDrive is a bottleneck, so I can't fault the G5 for that. I'd be curious to see how fast it would rip aiffs from the hard drive (not that anybody would pull aiffs from the CD and THEN rip them, but often times I create AAC files from AIFs created in CuBase).
These are about as harsh of criticisms as I can muster, which means the G5 appears to be a monster.
hulkaros
The Incredible...
He loads Mail, iChat, iMovie, iTunes, iPhoto, iCal, OmniGraffle, Transmit, Terminal and System Preferences. Safari and Elements are already open. The fun part is that they appear on screen in 10 secs tops and NOT that the icons stopped bouncing! Plus he is grabbing the video at the same time
Anyways, I think that the guy stated many times that he just "tried" to benchmark his G5 so any possible mistakes are fine with me!
I will let you know what I will find in the days ahead because my G5 is almost here!
Anyways, I think that the guy stated many times that he just "tried" to benchmark his G5 so any possible mistakes are fine with me!
I will let you know what I will find in the days ahead because my G5 is almost here!
Arden
Where mah "any" keys at?
8 out of those 10 applications are from Apple, and the other 2 are from Mac-loving companies. So it's no wonder that they load über-quickly. Now, if he tried to launch Photoshop, Lightwave, Final Cut Pro, Dreamweaver, etc. all at the same time...
Mindbend, you may get really quick rip speeds, but what quality settings are you using? Rob Griffiths used "MP3 format at 192kbps with VBR set to high" which is quite a high rate, and may also account for the slower rip speed.
Overall, I think this is a better test than any of those "official" tests. This is a real-world, out-of-the-box, day-to-day test of the system. He just puts it together, throws a couple programs onto it, and times it with a stopwatch. He doesn't optimize it for anything, set the processor to this or that, etc., much like a normal user wouldn't do. Considering what this machine can do without all that crap, now that is power.
Mindbend, you may get really quick rip speeds, but what quality settings are you using? Rob Griffiths used "MP3 format at 192kbps with VBR set to high" which is quite a high rate, and may also account for the slower rip speed.
Overall, I think this is a better test than any of those "official" tests. This is a real-world, out-of-the-box, day-to-day test of the system. He just puts it together, throws a couple programs onto it, and times it with a stopwatch. He doesn't optimize it for anything, set the processor to this or that, etc., much like a normal user wouldn't do. Considering what this machine can do without all that crap, now that is power.
DP 1 Gig G4 (original) 1 gig RAM
iTunes MP3 rip: 192KBS/High VBR
72 minutes of music took 6 minutes or exactly 12x with a 17x peak
-----------
The DP G5 peaked at 16x and averaged 11.6x
His SP G4 peaked at 8.5 and averaged 5.5x speed
This tells me, as I suspected, that the G5 is bottlenecked by the Superdrive. That's the only logical explanation. 4-5 minutes per average CD is still pretty quick though.
As for app launching, Photoshop has gotten slower and slower the last couple of versions at launching and general interface performance. Shame on Adobe. Final Cut takes longer than ever to launch, but runs better overall than previous versions. On a clean system, Lightwave launches instantly (less than a second). Fully ready to go. Amazing. I wish all apps could do that.
At a glance, I didn't see what OS he was using. I presume not Panther yet, but I could be wrong. As we all know, Panther alone will increase app launch speed on ALL machines, so in the end, the G5 app launch movie doesn't exactly blow me away, nice though it may be.
What does blow me away is the RAM-loaded Photoshop results, the initial Final Cut Pro feedback results (9 DV25 streams at once, holy lord!) and general "it just feels fast" reports. My G4 just feels sluggish all around.
iTunes MP3 rip: 192KBS/High VBR
72 minutes of music took 6 minutes or exactly 12x with a 17x peak
-----------
The DP G5 peaked at 16x and averaged 11.6x
His SP G4 peaked at 8.5 and averaged 5.5x speed
This tells me, as I suspected, that the G5 is bottlenecked by the Superdrive. That's the only logical explanation. 4-5 minutes per average CD is still pretty quick though.
As for app launching, Photoshop has gotten slower and slower the last couple of versions at launching and general interface performance. Shame on Adobe. Final Cut takes longer than ever to launch, but runs better overall than previous versions. On a clean system, Lightwave launches instantly (less than a second). Fully ready to go. Amazing. I wish all apps could do that.
At a glance, I didn't see what OS he was using. I presume not Panther yet, but I could be wrong. As we all know, Panther alone will increase app launch speed on ALL machines, so in the end, the G5 app launch movie doesn't exactly blow me away, nice though it may be.
What does blow me away is the RAM-loaded Photoshop results, the initial Final Cut Pro feedback results (9 DV25 streams at once, holy lord!) and general "it just feels fast" reports. My G4 just feels sluggish all around.