Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
PCMag's PowerMac G5 2.0GHz Review
- Thread starter MacBytes
- Start date
- Sort by reaction score
did any one else notice that they said that the G5 has 2 firewire 800 ports (one on the front) and only one firwire 400 port. i thought it was firewire 400 that had 2 ports (one on the front) not firewire 800. these little mistakes make me not like PC mag.
Re: Re: Re: How are you configuring this Dell???
The unshrap & wave tests were dramatically faster on the G5 vs G4 than the other tests, which is why I thought these two must have some speedup due to the G5 (more than just the MHz boost). The others were faster, but these were so much faster (unsharp almost 3x faster)
The unshrap & wave tests were dramatically faster on the G5 vs G4 than the other tests, which is why I thought these two must have some speedup due to the G5 (more than just the MHz boost). The others were faster, but these were so much faster (unsharp almost 3x faster)
Re: acrobat Vs. preview
I think they were testing Acrobat the pdf creation tool, not the viewer.
BTW, I use Acrobat Reader. Preview doesn't allow checking for embedded fonts and still has some printing glitches.
I think they were testing Acrobat the pdf creation tool, not the viewer.
BTW, I use Acrobat Reader. Preview doesn't allow checking for embedded fonts and still has some printing glitches.
Re: Hostile Aren't they
Cool. Just came in from a 30 miler myself on ye ole TCR Aero.
I think the most important part of the speed test was missed.
The Dell may have been about as fast as the mac, but WINDOWS SUCKS.
And OSX rox.
Just wait until Panther roars on the G5 Dual bad boy.
Cool. Just came in from a 30 miler myself on ye ole TCR Aero.
I think the most important part of the speed test was missed.
The Dell may have been about as fast as the mac, but WINDOWS SUCKS.
And OSX rox.
Just wait until Panther roars on the G5 Dual bad boy.
I guess it was an *alright* article... Especially coming from PCMag.
I think if anything, their 4/5 rating and performance testings just go to show how exceptional these high-end G5 boxes really are/will turn out to be...
The point for my post, however, is that PCMag would have done well to point out that a similarly configured Dell (priced for a full config. from Dell's site, just as the G5 was priced for a full config. from Apple's site) is actually a litte bit more *expensive* than the G5, and still lacks (among other things):
- PCI-X
- FW800
- SATA
- High(er) bandwidth system arch.
- *Quiet* fan operation
- Etc.
I.e, the G5 gives one far more bang for the buck (This is an Apple -- So it should be a big deal, as price was one thing that PCs generally trounced Apple on!), better build quality, (probably) better performance, etc.... and of course, that *sexy* case.
Then again, I guess it's called PC Magazine for a reason?
I think if anything, their 4/5 rating and performance testings just go to show how exceptional these high-end G5 boxes really are/will turn out to be...
The point for my post, however, is that PCMag would have done well to point out that a similarly configured Dell (priced for a full config. from Dell's site, just as the G5 was priced for a full config. from Apple's site) is actually a litte bit more *expensive* than the G5, and still lacks (among other things):
- PCI-X
- FW800
- SATA
- High(er) bandwidth system arch.
- *Quiet* fan operation
- Etc.
I.e, the G5 gives one far more bang for the buck (This is an Apple -- So it should be a big deal, as price was one thing that PCs generally trounced Apple on!), better build quality, (probably) better performance, etc.... and of course, that *sexy* case.
Then again, I guess it's called PC Magazine for a reason?
I'm probably wrong but I think when they refer to loading the 'controls' they are talking about the 'control subjects' i.e. opening the test files.
Anyone agree?
i_b_joshua
Anyone agree?
i_b_joshua
PS
That was an amazingly positive review for a PC mag. Credit to them for publishing it at all. Media silence hurts our favourite platform more than a few biased facts.
i_b_joshua
That was an amazingly positive review for a PC mag. Credit to them for publishing it at all. Media silence hurts our favourite platform more than a few biased facts.
i_b_joshua
wha wha?
Didn't the G5s on stage in WWDC just beat the crap out of the Xenon and P4? It just seems that it completely trounced them and now you see these tests and you feel like - "Ok. That's nice, it is on par. I now just feel ok. Even steven.". Was I watching the WWDC, too close to the RDF being emitted from my PB? Would a full load of RAM completely trounce these Win PCs? Or is it truly Panther being the wildcard in performance? Or Maybe...it was a 64 bit version of Panther in WWDC and all those apps in the WWDC were actually rewritten to take advantage of 64-bit addressing! A-ha! That is it.
I think I will get on my bike and roll down a hill...
Didn't the G5s on stage in WWDC just beat the crap out of the Xenon and P4? It just seems that it completely trounced them and now you see these tests and you feel like - "Ok. That's nice, it is on par. I now just feel ok. Even steven.". Was I watching the WWDC, too close to the RDF being emitted from my PB? Would a full load of RAM completely trounce these Win PCs? Or is it truly Panther being the wildcard in performance? Or Maybe...it was a 64 bit version of Panther in WWDC and all those apps in the WWDC were actually rewritten to take advantage of 64-bit addressing! A-ha! That is it.
I think I will get on my bike and roll down a hill...
Neat benchmarks, and good to see the G5 holding its own against a very high end computer... particularly since I finally ordered mine a couple hours ago (thanks in part to the edu discount, it looks like I'm paying over $1000 less than their test system for very similar specs).
I also found the fact that they were pitting it against a high-end Dell interesting since I was just configuring a similar CAD workstation for a client at the Dell site. Price and speed comparisons are always a shady area, but in general terms its pretty clear that the G5 can hold its own against a Dell workstation at a similar price. And it runs the MacOS. Score one for Apple.
Now, if you want detailed comparsions (hey, I'm a geek, I actually enjoy this stuff), the price comparisons are very difficult given Dell's highly confusing configurator (though it certainly does offer a wide variety of options). I played around with the numbers just for fun, and here's what I got along with a huge list of caveats:
I came up with $5,285 assuming they got:
-the 1MB L3 cache Xeons (probably)
-the cheapest 128MB graphics card Dell offers, the QuadroFX 500 (which is only $100 higher than the base 32MB card, and graphics cards wouldnt've made much difference in their speed tests)
-opted for the 2DIMM 2GB memory configuration instead of the four (same as Apple's)
-upgraded to a 120GB IDE HD
-got a DVD burner (cheapest)
(If you go with four DIMMS, you get a price $600 lower--$4,636, still higher than their Mac, but since the 2GB Apple Store configuration uses 2 1GB DIMMS, that's not exactly fair--you could just buy some cheaper ones elsewhere and upgrade them yourself.)
To get a similar configuration on both machines, you'd end up with a $5,767 Dell and a $4,958 Mac (explanation below).
Here are the issues I caught with this comparison:
-Dell kicks in a Palm Zire or cheap printer free
-The PC has a lot more internal expansion bays and two more non-PCIx slots (it does have 3 PCIx slots).
-The Dell has a 3-year warranty, which costs $250 from Apple (Applecare)
-The Dell has a built in U320 SCSI controller (on the motherboard?), which is worth quite a bit if you use it-- ~$250 value.
-The G5 has built-in SATA, and comes with a 160GB SATA drive, which Dell doesn't offer on that model. Adding another 40GB IDE drive for equivalent storage brings the price up $110, but a $50 SATA card and a $120 SATA drive tacked on to the base drive would be a better comparison.
-The G5 can go up to 8GB RAM instead of 4GB.
-The G5 comes standard with much better sound; you could add an $80 Sound Blaster Audigy 2 card to the Dell to bring it above the G5, if you wanted.
-The G5 has a FW800 port; the Dell only has 400Mbps ports (+$75 if you want the 1394b).
-The graphic card is the biggest one, since there are no workstation-class graphics on the Mac (bummer). Here's what I've been able to figure out:
The Radeon 9800 Pro is similar to ATI's workstation-class FireGL X2, but with half the RAM, a few less hardware tweaks that make the X2 better for a workstation, and a different driver (the consumer card is game tuned, the Workstation model is CAD/3D work tuned). The same goes for NVIDIA's Quadro FX cards vs. their high end gaming cards.
Comparitive benchmarks are tough (if not impossible), mainly since the cards are designed for different purposes, but if you want a very rough comparison, the 128MB 9800 Pro should be in the ballpark of an older FireGL X1, which (acording to Tom's) is a little slower than an FX 1000 but much faster than an FX 500. The X1 costs $300 more than the FX 500 from Dell.
I also found the fact that they were pitting it against a high-end Dell interesting since I was just configuring a similar CAD workstation for a client at the Dell site. Price and speed comparisons are always a shady area, but in general terms its pretty clear that the G5 can hold its own against a Dell workstation at a similar price. And it runs the MacOS. Score one for Apple.
Now, if you want detailed comparsions (hey, I'm a geek, I actually enjoy this stuff), the price comparisons are very difficult given Dell's highly confusing configurator (though it certainly does offer a wide variety of options). I played around with the numbers just for fun, and here's what I got along with a huge list of caveats:
I came up with $5,285 assuming they got:
-the 1MB L3 cache Xeons (probably)
-the cheapest 128MB graphics card Dell offers, the QuadroFX 500 (which is only $100 higher than the base 32MB card, and graphics cards wouldnt've made much difference in their speed tests)
-opted for the 2DIMM 2GB memory configuration instead of the four (same as Apple's)
-upgraded to a 120GB IDE HD
-got a DVD burner (cheapest)
(If you go with four DIMMS, you get a price $600 lower--$4,636, still higher than their Mac, but since the 2GB Apple Store configuration uses 2 1GB DIMMS, that's not exactly fair--you could just buy some cheaper ones elsewhere and upgrade them yourself.)
To get a similar configuration on both machines, you'd end up with a $5,767 Dell and a $4,958 Mac (explanation below).
Here are the issues I caught with this comparison:
-Dell kicks in a Palm Zire or cheap printer free
-The PC has a lot more internal expansion bays and two more non-PCIx slots (it does have 3 PCIx slots).
-The Dell has a 3-year warranty, which costs $250 from Apple (Applecare)
-The Dell has a built in U320 SCSI controller (on the motherboard?), which is worth quite a bit if you use it-- ~$250 value.
-The G5 has built-in SATA, and comes with a 160GB SATA drive, which Dell doesn't offer on that model. Adding another 40GB IDE drive for equivalent storage brings the price up $110, but a $50 SATA card and a $120 SATA drive tacked on to the base drive would be a better comparison.
-The G5 can go up to 8GB RAM instead of 4GB.
-The G5 comes standard with much better sound; you could add an $80 Sound Blaster Audigy 2 card to the Dell to bring it above the G5, if you wanted.
-The G5 has a FW800 port; the Dell only has 400Mbps ports (+$75 if you want the 1394b).
-The graphic card is the biggest one, since there are no workstation-class graphics on the Mac (bummer). Here's what I've been able to figure out:
The Radeon 9800 Pro is similar to ATI's workstation-class FireGL X2, but with half the RAM, a few less hardware tweaks that make the X2 better for a workstation, and a different driver (the consumer card is game tuned, the Workstation model is CAD/3D work tuned). The same goes for NVIDIA's Quadro FX cards vs. their high end gaming cards.
Comparitive benchmarks are tough (if not impossible), mainly since the cards are designed for different purposes, but if you want a very rough comparison, the 128MB 9800 Pro should be in the ballpark of an older FireGL X1, which (acording to Tom's) is a little slower than an FX 1000 but much faster than an FX 500. The X1 costs $300 more than the FX 500 from Dell.
Re: wha wha?
I think the wild card is the suite of tests being run. Time and time again, PC people can (and do) pick a different set of tests and "trounce" the Mac. What was important--and surprised the PC Mag--was that THEIR set of tests showed that the Mac was equal to or better than the PC. The first time in recent history that has EVER happened. For them to say it was (just about) equal is the highest praise!
I doubt a different config (more RAM, Panther) would significantly effect the results. Pather could help, but only to the degree that an app like Photoshop talks to it. For example, if Photoshop loads its data into memory and crunches it and that data crunching is the majority of ithe time, then Panther won't help much. What would probably help more is Adobe or Lightwave doing more optimization on their side. However, since the dual G4 beats the PC in the Acrobat tests, I would guess that means that the system is helping out. Since the system seems to be helping, I would guess that Panther would improve those scores even more.
In some tests, the Mac probably will never win, as those can be ones where MHz DOES matter. For all the talk that MHz doesn't matter, it can. If that PC's CPU is kepting crunching and there's no Altivec shortcut available, the Mac (in this case) would always be 50% slower (2 GHz vs 3 GHz). That doesn't happen often, which is why the G5 can run so close in performance.
I think the wild card is the suite of tests being run. Time and time again, PC people can (and do) pick a different set of tests and "trounce" the Mac. What was important--and surprised the PC Mag--was that THEIR set of tests showed that the Mac was equal to or better than the PC. The first time in recent history that has EVER happened. For them to say it was (just about) equal is the highest praise!
I doubt a different config (more RAM, Panther) would significantly effect the results. Pather could help, but only to the degree that an app like Photoshop talks to it. For example, if Photoshop loads its data into memory and crunches it and that data crunching is the majority of ithe time, then Panther won't help much. What would probably help more is Adobe or Lightwave doing more optimization on their side. However, since the dual G4 beats the PC in the Acrobat tests, I would guess that means that the system is helping out. Since the system seems to be helping, I would guess that Panther would improve those scores even more.
In some tests, the Mac probably will never win, as those can be ones where MHz DOES matter. For all the talk that MHz doesn't matter, it can. If that PC's CPU is kepting crunching and there's no Altivec shortcut available, the Mac (in this case) would always be 50% slower (2 GHz vs 3 GHz). That doesn't happen often, which is why the G5 can run so close in performance.
I'd wait 'till Panther comes out.According to some published sources and preliminary tests, it boasts overall system performance by 20-50 % on the g4,g3 class hardware. 
Plus there will be plethora of additional G5 specific optimizations.
Can't wait to see G5 shine on Black Cat
Plus there will be plethora of additional G5 specific optimizations.Can't wait to see G5 shine on Black Cat
Re: acrobat Vs. preview
I use Acrobat because it is, in fact, way faster than Preview.
Sorry, but in my world, Mr. Jobs is wrong on this one.
(tig)
I use Acrobat because it is, in fact, way faster than Preview.
Sorry, but in my world, Mr. Jobs is wrong on this one.
(tig)
There are plenty of people concerned with performance and usability of applications for specific purposes (e.g. their business) more than which OS or hardware they're running on. It's a "most-bang-for-the-buck" application issue, maybe even more-so with higher-end systems like the G5.
Correct me if I'm wrong, but Genie (since she mentioned it, not to single her out
) was considering getting a Dell instead of a G5 which I took to mean the OS really isn't that important for what she does.For how I use a computer the OS does still matter for me. I'd rather it could be transparent, but that's another topic...
Re: Re: Re: acrobat Vs. preview
Ah yes, well, my world is currently Jaguar. So, I'll reboot into Panther and see if it's any faster...
(tig)
Ah yes, well, my world is currently Jaguar. So, I'll reboot into Panther and see if it's any faster...
(tig)
Re: Re: wha wha?
It also depends on the type of data being crunched (and not just vector data). Both G5s and Xeons (and P4s) are capable of performing multiple commands in a cycle, depending.
(tig)
It also depends on the type of data being crunched (and not just vector data). Both G5s and Xeons (and P4s) are capable of performing multiple commands in a cycle, depending.
(tig)
I am extremely surprised that no one has talked about program optimization. Everyone knows that Adobe basically hates Apple (does everyone remember that "PC Preferred" web page?). Seriously, I hate tests on photoshop because you KNOW that Adobe puts WAY more effort into optimizing for PC. Yeah, they released the G5 plug-in. It's obvious that it does very little and they only did it to keep the mac customer's from complaining loudly. I'm positive if the program was evenly optimized that we'd see way different results.
I wish PC Mag would have mentioned that in their article.... *sigh*
I wish PC Mag would have mentioned that in their article.... *sigh*
Re: Re: wha wha?
The P4 pipeline takes about 3x as many clock cycles to do a single calculation, so if the program must wait on that calculation before preceeding, then it effectively gets cut to 1/3 the speed due to the pipeline.
Of couse the pipeline also does 3x as many calculations simultaneously, but unless it has around (21?) different calculations to do at the same time, it's not going to have all those steps full. Most likely the large pipeline creates a large amount of "bubbles" where there isnt any work being done. THIS EFFECTIVELY REDUCES THE CLOCK SPEED.
So what I'm saying is... yes, mhz does matter, but only if your processor is designed in such a way that it can use it. The P4 is not designed in a way to use all of it's power efficiently.
BTW, has anyone heard anything about the P5? The P4 has been out for a while now...
The P4 pipeline takes about 3x as many clock cycles to do a single calculation, so if the program must wait on that calculation before preceeding, then it effectively gets cut to 1/3 the speed due to the pipeline.
Of couse the pipeline also does 3x as many calculations simultaneously, but unless it has around (21?) different calculations to do at the same time, it's not going to have all those steps full. Most likely the large pipeline creates a large amount of "bubbles" where there isnt any work being done. THIS EFFECTIVELY REDUCES THE CLOCK SPEED.
So what I'm saying is... yes, mhz does matter, but only if your processor is designed in such a way that it can use it. The P4 is not designed in a way to use all of it's power efficiently.
BTW, has anyone heard anything about the P5? The P4 has been out for a while now...
prices are way off
this article says the dell is almost the same price.... wrong..price out a dell
workstation with dual xeons. with 2gb
of ram. a 146gb hard drive a better video card, a modem, without a monitor it comes to about 6k.
for that price you could put another 4gb
of ram in the mac and still save money.
this article says the dell is almost the same price.... wrong..price out a dell
workstation with dual xeons. with 2gb
of ram. a 146gb hard drive a better video card, a modem, without a monitor it comes to about 6k.
for that price you could put another 4gb
of ram in the mac and still save money.
They also said the G5 came in as heavier than the previous generation at 39 pounds. I believe when the G5 was introduced in June I looked up the weight of the G4 MDD and it was 41 pounds, so the G5 is actually lighter. Oh yeah, and thinner, than the previous generation.
Does anyone know if PC Mag ran the G5 with it optomized as far as the energy saver thingy? I don't have a G5 but wasn't there a post on here about G5's running faster with the manual setting rather than automatic, as they are sent from Apple?
Hey I'm tired and I don't know if the above is right but I think if Pc Mag really wanted a fair test they would have maxed out both machines ram, 8gig for the Mac and 4 gig for the PC and used a lot bigger graphic file
Not fair you say for the mac to have more ram, hey the pc has 2ghz more on the processsor end right?
daniel
Hey I'm tired and I don't know if the above is right but I think if Pc Mag really wanted a fair test they would have maxed out both machines ram, 8gig for the Mac and 4 gig for the PC and used a lot bigger graphic file
Not fair you say for the mac to have more ram, hey the pc has 2ghz more on the processsor end right?
daniel
Re: Look at the G4
I think the G5's performance was great. The Xeon system is a workstation, while the G5 is a regular personal computer. It might just be a marketing difference, but I still think it's interesting.
I noticed that, too. I'm pretty impressed that the G4 was able to beat the Xeon system on a couple things. At the price Apple is selling the Power Mac G4 for right now, I think it's a good value if you don't need the G5.Originally posted by otherguy
Did anyone notice those G4 times? The Dual G4 even smoked that Dell at several things...
I think the G5's performance was great. The Xeon system is a workstation, while the G5 is a regular personal computer. It might just be a marketing difference, but I still think it's interesting.
Re: Re: Look at the G4
The G5 is really a workstation in all but name. Look at the price and the features and the market, that "pro" market is a workstation market. But w/o the high-end graphic cards, it's not *really* a workstation. It's kind of an odd duck. 64-bit proc, monster bandwidth, SATA, etc, etc, but no FireGL, no quadro, no wildcat. The rest of the machine is clearly in the same realm as an SGI IMO. If only it had the 4MB L2.
The G5 is really a workstation in all but name. Look at the price and the features and the market, that "pro" market is a workstation market. But w/o the high-end graphic cards, it's not *really* a workstation. It's kind of an odd duck. 64-bit proc, monster bandwidth, SATA, etc, etc, but no FireGL, no quadro, no wildcat. The rest of the machine is clearly in the same realm as an SGI IMO. If only it had the 4MB L2.