PDA

View Full Version : MacWorld mag G5 speed tests


Ambrose Chapel
Sep 12, 2003, 07:13 AM
http://www.macworld.com/2003/09/reviews/macworldlabfirstg5testresults/

check it

Tiauguinho
Sep 12, 2003, 07:35 AM
Glad to know that MacWorld finally did some tests. The Dual G5 is the king! :)

Dont Hurt Me
Sep 12, 2003, 07:48 AM
still big gap from high end model to the singles, apple should make em all duals.

MacsRgr8
Sep 12, 2003, 09:06 AM
Exactly.

If you have a Dual 1 Gigger or faster G4, getting a Single G5 won't be a good investment...

the future
Sep 12, 2003, 09:18 AM
I actually find it quite impressing that the overall score of the single 1.6 G5 is higher than that of the dual 1.42 G4!

MacsRgr8
Sep 12, 2003, 09:52 AM
Hmmm.....
The Dual G4 1.42 is faster in iMovie Rendering, iTunes ripping, more FPS in Quake, faster in Photoshop, Cinema and MPEG encoding than the Single 1.6 GHz G5.
I assume that the "overall score" is a single processor benchmark.

mkaake
Sep 12, 2003, 10:58 AM
so now i suddenly have this urge to squander 3000 big ones that i don't have because i MUST have the dual 2.0...

of course, from the way it sounds, it would be about christmas before i got it.

but what a christmas gift.

matt

gwangung
Sep 12, 2003, 11:28 AM
Originally posted by mkaake
so now i suddenly have this urge to squander 3000 big ones that i don't have because i MUST have the dual 2.0...

of course, from the way it sounds, it would be about christmas before i got it.

but what a christmas gift.


Of course, by that time, they'll have announced dual 3.0s...

granex
Sep 12, 2003, 11:32 AM
Did anyone notice that the MPEG2 test of a 6:40 DVD occured in less than 6:00? This suggests that we are at the point of realtime full video processing on personal computers.

Physiognome
Sep 12, 2003, 11:35 AM
The 1.42 is slightly faster or equal to the 1.8 G5 in all the real tasks but Photoshop, as well. Also note the last paragraph about the Processor Performance setting.

Hopefully Panther will fix these numbers right up. :)

MacsRgr8
Sep 12, 2003, 12:09 PM
I think its good to see that the Dual 1.42 GHz G4's perform so well.
I don't really expect Panther to make such a huge difference on these tests.
Probably the new releases of Photoshop, iMovie and so on would make a greater difference.

Genie
Sep 12, 2003, 01:11 PM
Genie so happy!

Can't wait for mine to show up!

Abstract
Sep 12, 2003, 01:40 PM
Jezus, whoever bought a single proc PM got ripped!! Seriously, what a joke. They aren't worth the money. There's little improvement over a dual 1.42GHz, if any at all. :o I'd hold out for the duals, and ONLY the duals...

Maybe that's why Apple stopped selling their 1.42GHz systems and stuck with selling their 1.25GHz systems for a while. They knew what the reviews would reveal, and so they just too their 1.42GHz systems off their site.

Or maybe I forgot to take my ritalin today and am just rambling.....

MacsRgr8
Sep 12, 2003, 01:57 PM
Originally posted by Abstract

Maybe that's why Apple stopped selling their 1.42GHz systems and stuck with selling their 1.25GHz systems for a while. They knew what the reviews would reveal, and so they just too their 1.42GHz systems off their site.


There are lot of reasons why Apple stopped producing the 1.42 Ghz's....
Remember that the current 1.25 is based on the first rev. A of Dual 1.25 Ghz G4's - the one that can boot Mac OS 9....so:
reason # 1: Mac OS 9 boot, no FW 800
reason # 2: Not many 1.42 GHz procs around
reason # 3: 1.42 GHz is expensive
reason # 4: performance..... Not that the Dual 1.42 is much faster than the Dual 1.25, but as you can see the Dual G4's come too close (or even outpace) the Single G5

pgwalsh
Sep 12, 2003, 02:11 PM
I may have missed it, but was the OS optimized for the G5 in these tests? If not, then they'll need to be done again after optimization.

MacsRgr8
Sep 12, 2003, 02:17 PM
I already mentioned this, but I assume that the next releases of the software would be more of interest -G5 optimization speedwise- than the next OS release....
IMHO Photoshop 8 would be more "G5" optimized, than Photoshop 7 (w/ G5 "updater") running on Panther.

ZildjianKX
Sep 12, 2003, 02:26 PM
I can't believe they didn't post what videocard each model had in it. I mean, look at the quake 3 test... the dual G5 gets twice the FPS as the 1.8 GHz. They really should have tested them with the same videocards... otherwise that benchmark was pretty worthless.

So let me guess, the 1.6 and 1.8 GHz had the NVIDIA GeForce FX 5200 Ultra, and the 2.0 GHz had the ATI Radeon 9600 Pro? How is this a comparison between the models if they don't have the same videocards?

MacsRgr8
Sep 12, 2003, 02:32 PM
Originally posted by ZildjianKX
I can't believe they didn't post what videocard each model had in it. I mean, look at the quake 3 test... the dual G5 gets twice the FPS as the 1.8 GHz. They really should have tested them with the same videocards... otherwise that benchmark was pretty worthless.

So let me guess, the 1.6 and 1.8 GHz had the NVIDIA GeForce FX 5200 Ultra, and the 2.0 GHz had the ATI Radeon 9600 Pro? How is this a comparison between the models if they don't have the same videocards?

Well spotted.
Assuming they are testing the default configs, the Dual would have the 9600, and the Single the FX.... That is indeed a HUGE disadvantage for the single.
Good ad voor ATi, though ;)

ZildjianKX
Sep 12, 2003, 03:09 PM
Originally posted by MacsRgr8
Well spotted.
Assuming they are testing the default configs, the Dual would have the 9600, and the Single the FX.... That is indeed a HUGE disadvantage for the single.
Good ad voor ATi, though ;)

Maybe the people who performed the benchmarks at macworld.com didn't have the "technical know-how" to swap the videocards amongst the 3 G5s they had to do a proper test :D

The 1.6 GHz G5 also has 256 MB of RAM stock... it is still super crappy of them not to list the full specs of the machines they're testing.

I really want to see how a 1.8 GHz performs with a Radeon 9600.

MacsRgr8
Sep 12, 2003, 03:25 PM
Yeah, It is strange they didn't gave any details to exact configs.... It's the least they could mention. :rolleyes:

Ambrose Chapel
Sep 12, 2003, 03:31 PM
i'm sure when they publish the full reviews of the G5s they'll include all the info...they usually do.

Dont Hurt Me
Sep 12, 2003, 04:58 PM
i would guess they were all configured stock, so the singles have a fx5200 while the dual was a 9600 ati card. just my guess. the top machine is a screamer., but really how much do you need? iam happy has heck with a single g4 at 1.33 does everything i need.

daveL
Sep 12, 2003, 05:21 PM
Originally posted by ZildjianKX
Maybe the people who performed the benchmarks at macworld.com didn't have the "technical know-how" to swap the videocards amongst the 3 G5s they had to do a proper test :D

The 1.6 GHz G5 also has 256 MB of RAM stock... it is still super crappy of them not to list the full specs of the machines they're testing.

I really want to see how a 1.8 GHz performs with a Radeon 9600.
The article states that all systems had 512 MB of memory.

nospleen
Sep 12, 2003, 05:37 PM
Originally posted by Abstract
Jezus, whoever bought a single proc PM got ripped!! Seriously, what a joke. They aren't worth the money. There's little improvement over a dual 1.42GHz, if any at all. :o I'd hold out for the duals, and ONLY the duals...

Maybe that's why Apple stopped selling their 1.42GHz systems and stuck with selling their 1.25GHz systems for a while. They knew what the reviews would reveal, and so they just too their 1.42GHz systems off their site.

Or maybe I forgot to take my ritalin today and am just rambling.....

The 1.8 is on par with the 1.42 and costs 300 less than the dual 1.42. I think that would not constitute anyone getting ripped. Plus, once the apps are written for the G5, these stats should change. However , I may be a biased single processor G5 owner.:D

Mr Maui
Sep 12, 2003, 05:54 PM
For all of those who are whining that the Single processor G5s are a rip-off ... keep this in mind. Anyone who "upgraded" from a Dual 1.42 G4 ($3000 new) to a Single 1.8 or 1.6 G5 ($2400 and $2000 respectively) without seeing benchmarks first was foolish. :confused:

No one claimed that the Single G5s would be "significantly" faster than the Dual 1.42 G4s in all processes. However, for someone with an older computer ... a choice between single G5 or dual G4 would suggest that the Single G5 was a better buy (cheaper), a quieter machine, a faster bus, more memory capabilities (mind you all computers in benchmarks had 512MB ... the G5 can go to 4 GB ... which would improve performance significantly), a faster DVD burner, and will take ultimately take advantage of 64 bit processing when the software is available for it ... something the G4s can't claim. The G5 dual 2GHz for $3000 and the G4 Dual G4 1.42 GHz for $3000 a few months back ... was well worth the wait ... and the single G5s were a great replacement choice over the dual G4 1.42 for an upgrade option from something OLDER!!

Just my opinion! :D

The Shadow
Sep 12, 2003, 06:57 PM
I agree with you Mr Maui.

It'd be pretty unlikely that someone who paid full price for a 1.42 a few months back, if indeed they did want to upgrade, would go for the bottom of the range SP1.6.

So it's all a question of the end users circumstances.

Having said that, I've got an older G4533 and I am waiting to see if the mid range PM goes Dual next year (and whether Apple responds to criticism of only one optical drive) before I upgrade next year some time.

Cheers to all.

sjk
Sep 13, 2003, 12:45 AM
Originally posted by The Shadow
I agree with you Mr Maui. Me, too. :)

I wish benchmark results consistently provided more thorough descriptions of the testing conditions on each system. For these MW tests can we assume they were all run from a generic software environment without any other applications running besides the default Unix daemons in the background? Some baseline for comparison purposes is often incomplete or missing.

reflex
Sep 13, 2003, 02:34 AM
Originally posted by Abstract
Jezus, whoever bought a single proc PM got ripped!! Seriously, what a joke. They aren't worth the money. There's little improvement over a dual 1.42GHz, if any at all. :o I'd hold out for the duals, and ONLY the duals...

The one (arguably big) advantage that the single G5 holds over a dual G4 is that it's a 64bit cpu. Right now that doesn't mean much, but in the future there will be a 64bit version of OSX, which will probably take full advantage of it. The scores will likely be different then.

windwaves
Sep 13, 2003, 09:07 AM
If you have a dual G4 mac 1ghz or above, there is hardly any reason to upgrade to a single g5 mac. Only a Dual G5 2Ghz Mac would make sense.
So I am happy to see that as far as I am concerned my G4 Dual 1.25 piece of junk still holds its own, performance wise, v.s. the G5's except for the dual, their top of the line.

So, these g5 ... much hype about nothing, in my view, since only one of the new Mac really is a significant improvement over the previous PM G4 line up - again, on a performance basis. Possibily they have solved other issues, like the pathetic. ridiculous, never ending noise problems (with 999 fans in the box!).

Anyhow, lets see what Rev B brings us. I am ready.

windwaves
Sep 13, 2003, 10:26 AM
Originally posted by reflex
The one (arguably big) advantage that the single G5 holds over a dual G4 is that it's a 64bit cpu. Right now that doesn't mean much, but in the future there will be a 64bit version of OSX, which will probably take full advantage of it. The scores will likely be different then.

Of course, but these are hypothesis, nobody really knows/understands how this 64bit implementation will effectively translate into higher performance in real world apps. And so, it sounds to me completely meaningless to make a 64bit argument TODAY. What does it mean to say to someone who buys a computer: " the processor is 64bit architecture (wow !!!!), however there are no apps that really are aware of that; there will be in the future and we <expect> improved performance ???
I guess this means, if you must buy today of course it does not really matter; however, for those considering to upgrade from recent Macs, there is very little incentive to do so. Just wait and see.

The Shadow
Sep 13, 2003, 07:53 PM
Everyone seems to be making good points.

I would just add that it's all hypothetical to me because in addition to performance issues, I cannot upgade untill there is a PC emulator that will run on the G5.

Hopefully this will get solved soon.

Cheers.

Genie
Sep 13, 2003, 08:20 PM
Yeah- I agre the dual 2 is the only one I wanted to spend a bunch of money on right now.

I'd like to wait for Rev b, but who knows how long it will be before those actually ship, and whether the pricing will be as cool as the Dual2G is.

windwaves
Sep 13, 2003, 09:29 PM
I'd expect revisions to be quicker, Not that Apple won't do the same mistake again (announce when they don't even have a clue when they can ship), but it will be quicker. Again, it depends on your current system. My Ice Cracker (dual 1.25 FW800) will do for a while longer.