Slashdot Mirror
PC World: Apple G5 Gets Trounced By Athlon 64
StewedSquirrel writes "PC World magazine has published an article comparing the AMD Athlon 64 and Opteron versus Apple's G5 processor, both 64-bit contenders for the title of 'fastest desktop processor.' Apple has made many claims to be the first, fastest and only 64-bit processor for the desktop and workstation market, but (not mentioning the fact that Opteron beat the G5 to market by over 4 months) the benchmarks should speak for themselves. Of note is the 3.2GHz Pentium 4, coming in competitive with the G5, but significantly behind the Opteron and Athlon 64 systems."
last time i checked the operon was to be the server class amd64 cpu, where as the athlon64 was to be the desktop version.
if you're going to compare workstation class chips, compare the freaking workstation class chips...
From the article:
"But upgrading to XP 64 could mean giving up functionality without getting much in return. In fact, XP 64 looks like a throwback to Windows past: Its interface mirrors that of Windows 2000 or even Win 98. Microsoft has not disclosed what else will be in the OS, so it is possible that you'll still get most of XP's other features.
XP 64 won't have the 32-bit XP's support for DOS apps at all, nor will it run 16-bit apps (but it should have no trouble with 32-bit software). More important, 64-bit drivers for common hardware, such as printers, will be scarce when the OS debuts."
In moving from a Dual 1GHz G4 (Quicksilver 2002) to a Dual 2GHz G5, I have yet to find any software incompatibilities - everything works just fine.
This may change once my copy of Panther shows up, but my printer and other hardware continue to work for now.
Specialization is for insects. - R.A.H.
It might still get owned, but redoing the tests with the OS that the G5 was meant to run on will be a better comparison. What can it hurt, it's only 9 days away from release.
G5, Athlon64...any way you go it's an alternative to Intel. I think the importance isn't which is quicker but that they both offer serious alternative solutions to Intel which forces everyone to innovate. Both companies deserve credit for working toward better solutions for customers.
Of course we torture people, we need the information --Gen. Pinochet
It's still the fastest desktop processor, because there is no desktop OS that runs on the Opteron until Microsoft releases the XP version in 2004. And no, Linux is not a desktop OS - ie something regular people can use (yet).
I don't know why Apple shoots them selves in the foot with this speed BS anyway. Seriously I like my iBook for many reasons, but speed isn't one of them (because it's slow - although seems as fast as many PC laptops for some reason), but I'm willing to put up with a little drag to have a cool computer. They just aren't going to win the speed race and they need to realize that. They need to focus on the value of the overall computer where Apple is indeed ahead in many respects.
...you'll wake the Mac zealots!
Get paid to code OSS
So they compare 32-bit apps running on a 64-bit AMD chip to 32-bit apps running on a 64-bit G5 and conclude that the AMD chip is much faster than the G5.
This does nothing to benchmark the capabilities of the chips -- just the capability of the chips to run non-native apps.
Go back to your lives, citizens, nothing to see here...
Life is short: void the warranty.
"Our test suite, PC WorldBench 4, cannot run on Macs. The new Macs aren't great values either, as the top-of-the-line G5 ($3549 as configured) costs about $200 more than the similarly configured Alienware Aurora.
The dual-G5 sparkled in one main area: our Photoshop test, which it completed in 18 seconds, or about 17 percent faster than the Aurora's 21 seconds. The 1.8-GHz single-chip G5 ($2999) trailed at 27 seconds.
Elsewhere, the Alienware earned top marks, performing particularly well in the Premiere QuickTime test."
======
>>$200.00 is nothing and no direct testing comparision is funny.... This is pure marketing hype.
I think the jury's still out. We haven't seen OS's or applications optimized for either platform. However, both systems are still pretty damn fast. I think it's going to come down to what you like best. Personally, I like OSX better than Windows or Linux on the desktop. OSX gives me all the power and stability of Linux, and it's easier to use and prettier than Windows, and it runs Photoshop. I'm a photographer, so that's pretty important to me. I still run Linux on my servers, though...those Mac servers are ridiculously expensive.
We don't have a state-run media we have a media-run state.
You mean like the different flavors of Solaris, Linux, and BSD I have running on the machines here in my lab?
Unfortunately, convincing "management" to let me run anything but Windows on a "company" machine is an exercise in futility.
Specialization is for insects. - R.A.H.
It sounds simple, but in reality moving to an x86 platform doesn't make sense from a technical nor business stance.
First, it's not as simple as recompiling a few things at the "higher layers." All of Apple's partners would have to port their applications as well. Porting apps is not as simple as you'd like to think--see the OpenOffice port to OSX.
Second, moving to commodity hardware of x86 would turn Apple into just another software company. Apple very much is a hardware company and its the marriage of that hardware with exceptional software that makes their advocates voracious in their support.
AMD has nice stuff but if Apple were to use their processors they would be proprietary and for use by Apple only. The processor would be designed and built from the ground up for Apple--sharing next to nothing with AMD's other offerings.
So for now, let's just be happy that AMD and Apple both have cool stuff.
Of course we torture people, we need the information --Gen. Pinochet
Did you look at the apps they compared the G5s and the Athlon64s with?
/. calls this a trouncing?
Word- It's Microsoft, no shit it's going to be faster on windows, who would have guessed that?
Premiere - The video app that sucks so hard on mac that Adobe stoped making it. Try the same functions on FCP and watch it come out a few times faster.
Quake 3 - A game, 'cause you know macs are what everyone uses for gaming, and developers spend just as much time optimising their mac versions.
Photoshop - The only relavant and fair app they bothered to test, and the G5 is noticablly faster than any of the Athlon 64 systems, beaten only by the Opteron.
And
"The worst tyrannies were the ones where a governance required its own logic on every embedded node." - Vernor Vinge
I cannot believe this is taken seriously on Slashdot.
Let's look into this more closely: the PCWorld team tested only four applications, one being Microsoft Word, FFS, and another being Premiere, which is no more supported on the Mac, runs in Classic and is leagues behind Final Cut Pro in terms of performance, as anyone with a clue in Mac video processing will tell you. This alone qualifies this comparison as biased in my book.
Where is the After Effects test ? And where is the Mathematica test ? Did you only know that any G5 will trounce an Athlon 64 in these apps ?
Also, looking at the results, I can hardly call it "trouncing the Mac". Only one in the four apps make use of the 2GHz' second CPU (Photoshop), and dutifully the G5 beats the PC in this test, and the scores in the other tests (not counting the Premiere's joke of an application) are not even that far apart.
Lies, damn lies, statistics, advertisements and benchmarks.
Aside from benchmarking Word for Mac against Word for Windows of all things, what does this actually prove? That Macs don't run software as well as Windows does when it comes to software that has been available for Windows longer? I'd be more interested in a price comparison between the systems.
No software-RAID setup on the Mac? Why RAID on the other machines?
Seems kind of one-sided.
I've argued with benchmarkers over and over about this, Premiere is a lousy benchmark, used only by people who want to stack the deck against Macs. Premiere is highly optimized for PCs, and highly unoptimized against Macs. Fortunately that benchmark will go away soon since there won't BE any further Mac Premiere versions.
If you want to do a proper test, you'd use a crossplatform product that runs equally well on both platforms and is highly optimized for dual processors, like Discreet Cleaner or Combustion.
There's only one benchmark I can think of that is more worthless than Premiere, the "MSWord scroll test." For some stupid reason, some benchmarkers think it's a useful test to see how fast the can scroll to the end of a long Word document with the arrow key. Unfortunately, Word has a delay loop built into the scroll function, it even changes the delay loop depending on the speed of the CPU. The results are totally useless.
both 64-bit contenders
Both the G5 and the AMD64 are great chips, but they really only represent the intrustion of 64 bit computing in the popular consciousness, not the actual beginning of 64 bit computing.
Compare their performance with the last Alpha chip, development of which was cut off years ago, and tell me again how the best is being brought to us.
Even as Intel picks the carcass of Alpha to revive the still-born Itanium series, the killed off Alpha chip line has performance that embarrasses HP into covering it up.
"Provided by the management for your protection."
Compare the single G5's results with the dual G5's: only Photoshop makes use of the second CPU.
So they're basically pitting ONE G5 against ONE Athlon64 in the other tests.
But this is not the only incoherence with this test. Using a Classic, unsupported application like Premiere instead of the native After Effects ? Testing a Mac's performance in Microsoft Office ? This is a joke.
not to mention they picked apps that run like crap on Macs...or don't have an OS X port (premier).
out side Photoshop (which the Mac won hands down) the rest of the apps are plain stupid choices.
I am the Alpha and the Omega-3
rock-solid, sure... Linux has been stone stable since before 1997. i'm talking about look and feel.
every year it's, "Linux on the desktop has made huge improvements in the last year," but it never quite gets there.
use OS X for a week or two and you'll see what i mean.
See: http://www.digitalvideoediting.com/2003/06_jun/fea tures/cw_macg5_interview.htm. Funny stuff.
[excerpt:]
DMN: Now, you're saying it's the first 64-bit desktop machine. But isn't there an Opteron dual-processor machine? It shipped on June 4th. BOXX Technologies shipped it. It has an Opteron 244 in it.
Rubinstein: Uh...
Akrout: It's not a desktop.
DMN: That's a desktop unit.
Akrout: It depends on what you call a desktop, now.
---
S
Hmm...
So this is how we benchmark two different platforms these days?
For everyone's information, I should not have to point the following out, but here we go... the benchmarks were taken from the following apps -
Quake III, developed on, and for, x86 over 5 year period of programming research and enhancement. Later ported to OSX in a week by OmniGroup.
Word, developed on, and for, x86, by the developer who also wrote the operating system running on the PC's. Ported by MBU to OSX.
Photoshop, Adobe develops Photoshop in a very balanced way for the two platforms, and these are the results for this test -
Fastest 50MB image = 17 seconds, G5 = 18 seconds
Fastest 150 MB image = 47 seconds, G5 = 51 seconds
The final test was a Premiere rendering, where almost all the systems tested did the job in 3 or 4 seconds. The fastest was 3 seconds, the G5 did it in 4. This is Premiere which no longer exists as a current ongoing product for OSX.
Does anyone see just how biased and unscientific this all is?
Oh, and I didn't mention that most of th PC's had double the graphics memory, and had RAID as their primary storage.
This article is FUD.
-Nex
This sig has been deprecated.
Offtopic? No. Redundant? Yes.
What's your point? That just means the G5 is a versatile chip. The person who submitted the article stated that the Opteron beat the G5 to market by 4 months..well, so what? There have been servers running 64-bit processors for a long, long time. Opteron is nothing new in that respect (its hybrid design is admirable, but a necessity, now that most 64-bit CPUs carry them). The G5 however, was marketed as a desktop solution, and was sold as one. The Opteron just was not. Thus, the time scale is completely negligible.
Also, the 970 may be a server chip now, but again you're missing the point: Opteron was never meant for the desktop.
--- What
We're all geeks, so we love benchmarks.
It's too bad that no rigor is applied to 99% of the benchmarks that are applied.
Raw CPU benchmarks like SPEC end up being compiler tests rather than processor tests.
"Real-world" app tests like this one are better, but only if the apps used are representative of apps used by the person reading the benchmark. They are not a realistic measure of holistic system performance.
Adobe Premiere? Come on! Does anyone on the Mac use that at all anymore? Is it even OSX native? Since Final Cut 1.0 came out 2+ years ago, anyone who considered that dog Premiere deserves what they get. Isn't there a better editing package on the PC, or is the Mac just that much better for video editing?
Microsoft products should be excluded from benchmarks on Microsoft's OS. Of COURSE Microsoft optimizes performance of their apps on their OS more than their apps on other OSs. That test is pointless.
The Quake test would be valid, except as many people here have pointed out, it's a 32-bit app, so it's not using any of the 64-bit capabilities of these boxes. When we get a native, 64-bit version and can compare it to two boxes with the same ATI video card, then it will be a valid test.
- Vincit qui patitur.
Those jokes are so freakin old they don't even touch funny.
But anyway, is this really a fair comparison? A lot of the performance that they measured has to do with how the software is written and compiled.
I just sent this letter in to PC World. I think it pretty much covers all the mistakes they made in the cross-platform benchmarks.
>--
I have been a long time reader of PC World, and have much respect for your magazine. However, I am yet to see a more abject review than the "64-Bit Takes Off" what was presented in your November 2003 Edition.
Let's start with the choice of Microsoft Word. Undoubtedly a widely used piece of software, and Microsoft incredibly allowed Office v.X for the Mac to receive a number of features that the Windows version is yet to receive. There is, however, one thing that Microsoft will not allow Office for the Mac to achieve; and that is performance parity. To add to this, much of the codebase of Office v.X is left over from the good ol' days of MacOS 9 - reflected in the fact that Office is still a Carbon app. So, although Office on the Mac is extremely widely used, it's of dubious use as a means of comparing performance between processors. Unless, of course, all you do is Office and it's not presently running fast enough for you.
Next. Premiere. This is what stunned me. There is a reason that Premiere doesn't work very well on the Mac. This is because absolutely nobody who does video editing on a Mac uses it. Period. Final Cut Pro wipes to floor with it; not only in functionality, but performance also. Of all the ways you chose to benchmark the G5s, this surprised me the most.
In the Quake test, the Mac was hamstrung by the fact that it only had a 128MB video card in it. I also may be wrong in making the assertion, but doesn't the 256MB ATI 9800 Pro run at a faster clock rate than its 128MB cousin? This would account for quite a performance differential. Despite the fact that Macs aren't really known for games, no other computer with a 128MB graphics card beat it.
The next test was Photoshop. This is the one app you benchmarked in which some 64-bit optimisations have taken place for the Mac, and is also an app that many people use on the Apple platform. In this test, the G5 beat everything on offer from the x86 world by quite a handy margin.
What makes this even more impressive is that the G5 system you benchmarked is running on a stop-gap operating system release from Apple. OS X 10.3, codename Panther, has been specifically designed to take advantage of the G5's 64-bit CPU structure; it's out in barely a week.
I would certainly be interested to see a re-run of the tests, if you think that this feedback is valid. Cross-platform benchmarks are notorious for being difficult to standardise; I do, however, believe that if done properly they can be both useful and interesting.
-- james
Well, I agree, but you're making the wrong impression. x86-64 and the G5 BOTH utilize 32bit instructions; they merely add on new instructions. So their ability to run 32bit code is paramount.. Plus not every part of every program needs 32bitness (for same reason why we still have 8bit instructions (e.g. character manipulation) or 16bit instructions (wanting integer modulation)).
Also, there are proven benchmarks that show that the opteron indeed runs faster, though due to the many differences between the opteron and the K7/P4, it's hard to narrow down exactly where the performance increase derives. The key feature, as stated is the doubling of the number of registers... The x86 has NEVER been able to perform register optimizations (like loop unrolling, etc) because there just arn't enough registers for even trivial book-keeping operations. The only way of really speeding up such apps is to have smart CPU's which rename explicit registers into remappable internal registers while instructions are running in parallel.. It's very hard however (and EXTREMELY cpu-specific) to have the compiler write code that can take advantage of this.
So having more regs means at least all the book-keeping can be explicitly stated by the compiler, and MANY explicit memory load/stores can be avoided (including CISC: add [mem], [mem], [mem] operations).
BUT, here's the misleading part.. The G5 has been RISC for a LONG time.. And one of the key tenents is 32+ registers. So while the x86-64 nicely ups the anti to a whopping 16 registers, don't get too excited, we're still living in the dark ages.
The main reason is that we're still doing register renaming, so to have 32 explicitly named regs you'd have to have nearly a thousand renamable registers.. This means slower addressing time (more propagation levels for a greater power-of two number of addressable registers). Is this additional cost warrented? The alternative is removing the power of the existing renaming facility, which may hurt more than help. So, given that CISC operations can still efficiently use memory addresses as virtual registers (e.g. the CPU keeps them in high-speed caches), it's a cheaper / faster over-all solution to just stick with 16 registers.
It still doesn't solve the problem of loop-unrolling however, so I'm not particularly happy with the decision (Being a compiler whore myself). AMD's not going to get another opportunity to radically alter the x86 line.. Intel is going to do this next and AMD will have to play catch up again.. And it's possible that the x86 won't last another major iteration (e.g. x86-64-AI?)
-Michael
Apple never said first 64-bit workstation. Only first 64-bit personal computer. Get it straight.
The new Macs aren't great values either, as the top-of-the-line G5 ($3549 as configured) costs about $200 more than the similarly configured Alienware Aurora. (Prices do not include a monitor or speakers.)
Ah, the well known high quality brand Alienware, of course Apple doesn't have a fighting chance against a brandname like that. Alienware's years of service in the Fortune 500, the constant stream of quality awards in consumer magazines. Apple's crazy to ask 200 dollars more, the G5 should be half the price to even be considered.
And the speedtest, well, we all know everybody is still stuck on the Mac with Premiere as their main application running in Classic. Why didn't they test Netscape Navigator 3.0 as well? I mean, that's todays most popular webbrowser!
The POINT is that Apple never marketed the G5 as the fastest workstation. All Apple marketed the G5 as was the a) first 64-bit desktop (and if your definition of desktop differs from "a pre-built box from a well-known company that an ordinary human might buy", that's your problem, not Apple's), and b) the fastest desktop around at the time.
Saying, "Ooh! Ooh! New computers have come out! There are benchmarks against computers Apple wasn't talking about! The G5's not the fastest! Apple LIED!" is just plain dumb. Of course faster computers will come out! Apple isn't dumb enough to think or claim that their first-generation G5s will always be the fastest, and anyone who thinks they were claiming that is dumb.
And does anyone else see the possible conflict of interest with PC World running these benchmarks? Now, note that I'm an Apple fan. However, I won't completely believe any benchmarks that are done by anyone with an interest in seeing either side win. And it would probably be best if both computers were running something neutral, like a Linux or a BSD. Does anyone really believe these benchmarks are any more fair and unbiased than Apple's own???
Dan Aris
Fun. Free. Online. RPG. BattleMaster.
I am seriously considering getting a G5 just to run Matlab simulations. Where's the tests of something that may stress the hardware a bit, like Mathematica or Matlab?
These benchmarks are a bad joke. My pentium II or Athlon box runs Word pretty fast.
..don't panic
SPECfp base2000
2Ghz G5 - 840
Opteron 146 (2Ghz) - 1291
SPECint base2000
2Ghz G5 - 800
Opteron 146 (2Ghz) - 1170
SPECfp rate2000
Dual 2Ghz G5 - 15.7
RackSaver RSN-1164/op (1.8 GHz Opteron) - 22.5
SPECint rate2000
Dual 2Ghz G5 - 17.2
RackSaver RSN-1164/op (1.8 GHz Opteron) - 24.0
These numbers seem to back up the PCWorld tests.
I'd agree with you, accept the Office isn't exactly what I'd call a high performance app, no one that buying any of these machines is buying them cause they need better performance out of MS office. Now a comparison of the most commonly used programs on each machine is fine, but keep in mind the target audience of these machines.
----- Question authority, but not ours. Hate the man, but we're not him.
Their tests are largely I/O-bound and video card related, too. It's a system comparison, not a processor comparison. If you have different I/O or video card, you'll get different results.
So who has the faster processor? Who knows. I suggest you buy the system you like the most.
I had a chance to try a 3.2 GHz P4, an Opteron, a dual G5, a 3.06 GHz dual Xeon, and an Athlon 64, using a suite of productivity and multimedia apps. Here's the verdict:
The P4: Very very fast.
Opteron: Super fast.
Dual G5: Really really fast.
Athlon 64: Totally fast.
Dual Xeon: Nice 'n fast.
Telling results! Unfortunately since I have put so much effort into accurate, impartial analysis of the test results, and participating in all the arguments with disbelievers and naysayers, I have not had a chance to get any work done for months. But who wants to use CPUs for productive tasks anyway, when it is so much fun to sit back and watch them "trounce" each other!
Also, the modern x86 processors have logic to help ease the problem of a limited register set. It's called register renaming. I am not sure the extra registers will have a profound effect on the performance.
There seem to be lots of complaints about the fairness of the benchmarks.
If you want to compare the performance of CPUs with different ISAs, then you need to have the benchmark source code to compile it to the target ISA. This also brings the performance of the compiler into question but there's just no other way to do it if you want to compare CPUs with different ISAs.
One benchmark that people in the industry use to measure relative CPU performance across different ISAs is the SPEC benchmark. Just about every single computer maker from Dell to HP to Sun have submitted scores. Apple has not. This is in an of itself very telling. What is Apple afraid of? I'm hoping IBM releases a computer based on the G5 so that we can get some idea of the SPEC performance. Apple seems to believe it has something to lose by submitting a SPEC score.
Don't like SPEC? Please suggest some other CPU intensive applications to benchmark that have source code and publish your results.
When reading my analysis, bear in mind that an average person with a stopwatch has +/- 1 second margin of error per test, so anything within two seconds is considered the same time.
Also consider that most machines spin down their hard drives when not in use, leading to up to a five second stall. Because there was no aggregation of multiple tests, tossing out any outliers in the process, these test results are basically useless, but you can consider them to have a +/- 12 second margin of error.
Finally, bear in mind that my analysis is extremely biased. Please look at the facts yourself and make your own decisions. Do not blindly accept my opinion as truth, as doing so doesn't do anyone any good.
Analysis of results:
- Render test: all times identical.
- Quicktime test:invalid (see below).
- Photoshop 50 MB test: tie between Polywell 2 and dual G5 for first place
- Photoshop 150 MB test: tie between Polywell 2 and dual G5.
- Quake tests: invalid.
- MS Word tests: invalid.
Reasons for invalidation of Quicktime test:- If two machines with similar performance suddenly show more than a factor of two difference, this almost always means that only one processor is being used on the slower system due to differences in the software used.
- The test is poorly described so that it cannot be reproduced. There is no "Quicktime format". Quicktime is a wrapper movie for any of dozens of formats. QuickTime has different default codecs in different versions. I doubt Premiere installs the same version of QuickTime that most Mac OS X 10.2 users would have installed (thanks to Software Update), thus there is a good chance they were using different codecs in this test.
- According to the QuickTime API docs, your application has to be modified to take full advantage of multiple processors when compressing images. Since Premiere for the Mac was last updated not long after that support was added in QuicktTime (as best I can tell), odds are very good that it does not use the new APIs, while recent Windows versions almost certainly do.
Reasons for invalidation of MS Word test: factor of four difference clearly indicates that software is not comparable across the two platforms. The results are beyond any sane person's ability to accept from nearly equivalent machines running even remotely similar code.Reason for Quake test invalidation: this should be dependent on graphics card performance, not CPU performance. The G5 beat all but one configuration with an equivalent video card. This one configuration inexplicably was about 50% faster than all the other configurations. Since at least one machine in each 128M speed class uses 8x AGP, it is safe to assume that there are substantially different versions of ATI's drivers being used in these tests, rendering any results meaningless in terms of the performance of the machine itself. The most likely (but hard to prove) interpretation of these results is that the G5 performs slightly better than any Athlon64 when given an equivalent video card, and that the one machine is either mislabeled or has a newer version of the ATI drivers than the G5 and the other 128MB PCs.
But that's just my opinion. I could be wrong.
120 character sigs suck. Make it 250.