Analysts and others are raising questions over Apple Computer’s claim that its new Power Mac G5 is the world’s fastest personal computer. Articles at News.com, ExtremeTech. Apple is defending their benchmarks. In the meantime, the first reference of how faster the G5 is against the G4s, is available.
Apple Benchmarketing Raises Questions
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Once an independent agency gets its hands on a G5 machine, test it heads-up with a Dell. Only, you should test three times…..
– Running Linux on both machines
– Running NetBSD on both machines
and
– Running OSX on the G5 with the apple-optimized compiler, and WinXP on the Dell using one of the 1337 compilers.
The first two should show quite a bit by themselves.
Of course, all of this is grandstanding. We might as well go by BogoMIPS. CPU speed is only a part of overall system speed. When you’re talking differences of less than 10%, you can make that up in other areas (i.e. disk subsystem).
I suggest Shake Linux and Shake OSX tests and Maya rendering on all 3 OSes. I agree with photoshop tests, and it would be a good idea to include a couple filter tests too. How about a little gaming test too, even if its old like Q3. The gpu’s are available for both. Who cares about optimized, optimized = realworld for both sides. Yes I said on both sides.
that some of the XBench CPU results are lower than the results of my dual 1GHz G4. Only floatingpoint is significant faster. Looks like I am still happy with my current machine.
Maybe the results get better when XBench ist copiled for 64Bits.
But I think I wait for an upgrade decission when the reach the promissed 3GHz.
People seem to live in a fairyland where there is The One Benchmark to determine The Fastest Machine.
And what was Jobs supposed to say – ‘the worlds fastest, or maybe second fastest, personal computer, as long as AMD doesn’t one-up us before August’?
Get real.
Jobs ALWAYS says Apple is the best or fastest. What has most PC people up in arms is how close it actually is. And infact are trying to brand Apple as cheating, when infact ALL PC manufactures tend to tweak things. But Apple this time used a trird party. So if any results are skewed it should be first taken with them. They are PAID to give results that are accurate. And instead of flogging APPLE maybe some people should be flogging the company that performed the tests.
After all from what I understand Apple has the only VERIFED (what ever that means) test results.
Last I have a distinct feeling that when the G5 DUAL is availible with Panther, The difference between PC’s and MACs for realworld will not be as far apart. And that is why I think people are screaming
I think these articles are plain and simple flame bait.
So, the unfair bench marks by the pc makers are now compared to unfair benchmarks of Apple ? I’m outraged. Not..
Fact: A G5 is a massive improvement over the G4
Fact: The new bus arch is a massive improvement over the G4
Fact: MacOS is the most rapidly evolution of any major OS
Fact: 2999 for a dual 2Ghz G5 is competitive with commodoty PC crap
Fact: you want one if you can afford it, or you’re just not enlightened yet.
To me benchmarks dont mean a thing, Sure in some things it may be faster but not all of them. Second I take it with a grain of salt because I asked for specs on the Dells and was not given an answer. I wont buy it because of the price basically and I think that is what the main problem will be with most consumers. So far tho WWDC has been interesting.
Looking at all the information on the table, Apple did a good job on the benchmarks.
If we want to get into the “expose” of benchmarks, why not take a gander at:
AMD’s XP 3200… which clearly does not perform up to its “3200” label.
All of Sun’s computers whose SPEC ratings have been lagging for years and still Sun’s ads claim Sun makes a performance processors.
All the database benchmarks, TPC and such, that are incredibly hacked and tuned just for the test.
The recent nasty between FutureMark and Nvidia for “cheating”… which later got changed to “optimization”.
And if we are going to start looking into benchmarks, why not accounting practices?
Let’s get back to talking about real stuff, not slimy attacks on Apple for doing a GOOD job with their benchmarks!
there are statements in there that infact Apple and test company tried to make a fair comparision. And in fact, the effort might not been that far off, time will only tell. Once some independents get the computers and test them.
….and benchmarks.
Seriously the FeeCee fan boys hace gone into some serious FUD overdrive. So the G5 is not that bad of a system, deal with it….
If a Mac with a G5 running OSX is as fast/smooth as my Dell P4 2.8 ghz running optimized Windows XP (or Longhorn, I might just have to buy one
But I don’t think you really need benchmarks to know if it ‘feels’ fast, if that makes any sense.
Well, its great to see that Apple and the PPC world finally have a decent desktop system. Hats off to Apple. I’ve watched the video of the keynote, nice.
The benchmarking issue isn’t really that important. Apple finally have a machine on par with the best AMD and Intel can supply. The bandwidth of the G5 is nice, and a PPC architecture with its numorous registers (much more than x86 legacy) give some apps a real boost. Those who need 64 bits will love the box. The price is sweet, and its great to see a big box manufacturer take care of the heat and noise issue in modern PC’s. Apple have done well.
6 months ago I purchased a revised system, so I wont be considering the G5 now (already have an XP2400 overclocked). If however I need to buy a new machine in September, I’d be very tempted to get the G5.
Live and let live.
A computer that benchmarks slow but runs fast when I use it than one that benchmarks fast and runs like a dog.
Benchmarks are a tool for determining how fast a system might do in the real world. I use real world applications to do work with not benchmarks.
P.S. if someone gives me a G5 machine I’ll say it benchmarks faster
Windows XP:
How fast does it upload your information to Microsoft?
Windows XP:
How many innocent people get put in prison for each copy of Microsoft Windows that Microsoft sells?
Windows XP:
How much pirated code does it contain? Why does Microsoft argue so vehemently for closed source?
I think the G5 looks great. While it may not be the fastest computer available it is the only 64 bit desktop available right now. It also looks like a great improvement over the G4.
However, I still think Apple is losing bigtime in price/performance comparison. I really don’t see why Steve Jobs is content with leaving Apple in its little niche when it could be so much more.
Oh well, I can’t currently, and probably never will be able to justify the expense of a Macintosh. It just isn’t worth it when I can get free OS’s and fast, cheap hardware.
The SPEC part of it. All these critics are ignoring the app part. Go watch the video. Dual 970 crushed dual xeon over and over on photoshop, mathematica, etc. It was embarrassing for the xeon, frankly.
…I’ll rebroadcast this note, from the News.com article linked to:
“Peter Glaskowsky, editor-in-chief of Microprocessor Report, said a company could get better benchmark results using a Dell machine with Intel and Microsoft compilers than with a Linux machine and GCC compiler. However, he also noted that Intel’s chips perform disproportionately well on SPEC’s tests because Intel has optimized its compiler for such tests.”
In other words, ICC is optimized specifically to do well on those benchmarks. This doesn’t violate SPEC’s rules–but it casts a certain level of doubt on the claims of ICC’s incredible superiority over GCC to begin with, and–unless it’s found that Apple specifically optimized GCC for the SPEC tests on the G5, something no one has claimed–it definitely weakens the arguments against having used GCC on both platforms for the comparison. (Either you use a compiler designed to explicitly tweak for those tests on both machines, or you use one that doesn’t have benchmark-specific tweaks on both.)
the tests were fair. there is no penalty.
move on.
http://www.apple.com/powermac/performance/
http://www.veritest.com/
http://www.veritest.com/clients/reports/apple/apple_performance.pdf
Greg Joswiak, vice president of hardware product marketing at Apple, in a phone interview today, defended Apple’s performance claims for its upcoming Power Mac G5, after they came under fire in the wake of yesterday’s announcement.
Joswiak went over the points in turn, but first said that they set out from the beginning to do a fair and even comparison, which is why they used an independent lab and provided full disclosure of the methods used in the tests, which would be “a silly way to do things” if Apple were intending to be deceptive.
He said Veritest used gcc for both platforms, instead of Intel’s compiler, simply because the benchmarks measure two things at the same time: compiler, and hardware. To test the hardware alone, you must normalize the compiler out of the equation — using the same version and similar settings — and, if anything, Joswiak said, gcc has been available on the Intel platform for a lot longer and is more optimized for Intel than for PowerPC.
He conceded readily that the Dell numbers would be higher with the Intel compiler, but that the Apple numbers could be higher with a different compiler too.
Joswiak added that in the Intel modifications for the tests, they chose the option that provided higher scores for the Intel machine, not lower. The scores were higher under Linux than under Windows, and in the rate test, the scores were higher with hyperthreading disabled than enabled. He also said they would be happy to do the tests on Windows and with hyperthreading enabled, if people wanted it, as it would only make the G5 look better.
In the G5 modifications, they were made because shipping systems will have those options available. For example, memory read bypass was turned on, for even though it is not on by default in the tested prototypes, it will be on by default for the shipping systems. Software-based prefetching was turned off and a high-performance malloc was used because those options will be available on the shipping systems (Joswiak did not know whether this malloc, which is faster but less memory efficient, will be the default in the shipping systems).
As to not using SSE2, Joswiak said they enabled the correct flags for it, as documented on the gcc web site, so that SSE2 was enabled (the Veritest report lists the options used for each test, which appears to include the appropriate flags).
Greg Joswiak, vice president of hardware product MARKETING at Apple
Oops.. OS News didn’t forget… the preceeding interview is in there…
“think about htis one long and hard………Greg Joswiak, vice president of hardware product MARKETING at Apple”
Ya, you’re right… Thats proof enough that Apple lied… [roll eyes]
The disputed claim was that Apple somehoe dumbed down the Intel machine and revved up the G5.
Not only did Apple not dumb down the machine, they put it in a situation that would achieve the best results for a fair comparison.
Its VERY important that you realize that he said that the Dell numbers would be higher with the Intel compiler, but that the Apple numbers could be higher with a different compiler too. They used the GCC compiler SPECIFICALLY because it was the only way to guarantee fair results.
The results would be inconclusive if different compilers were used.
…yet would probably receive the same results.
The specs ARE accurate. Get over it.
Its VERY important that you realize that he said that the Dell numbers would be higher with the Intel compiler, but that the Apple numbers could be higher with a different compiler too. They used the GCC compiler SPECIFICALLY because it was the only way to guarantee fair results.
By the way, the intel compiler is really optimised for spec tests. It’s weird that no one said anything about Intel cheating on benchmarks 🙂 Just look at the results of the Spec test and compare it to GCC. It is way out of proportion if you compare the applications compiled with GCC versus ICC.
But hey, today Apple is evil and Intel is good, I guess. I wonder what it will be tomorrow.
Apple uses Gcc because it can be sure that it is not optimized for the spec tests. Gcc is rather optimized for the plateform, like pentium 4 or now the G5.
But every developer knows that the compiler ICC of Intel is optimized to give a big score with the spec2000 benchnark. And its not only intel, amd does it, and Ibm does it. If you take the compiler of Ibm and run the spec2000 tests you will get higher numbers than apple, because IBM tries as well to optimize their compiler to this kind of benchmark which is very popular and so every compagny has to make a big score.
I really don’t think that apple’s result are unfair foe the p4, sse2 was enable amd the hyperthreading as well (not for the spec_rate) during the tests.
And when you see the demo of the keynote you see that on photoshop, luxologic and mathematica the Xeon pc is explosed by the G5 and disintegrated in the audio test.
that after reading yesterdays Troll-fest that it wouldn’t have mattered WHAT Apple did to make the tests as level as possible.
If Apple outperforms an x86 you can bank on the fact that the PC fanboys will cry shenanagens.
1. AMD is toast if they can’t get to 90 nm by Q2 04.
2. If Intel ships Prescott in Q4 03, signalling the ramp up of their 90 nm, watch out. That’s plenty of clock rate scaling (4+ GHz) to fend off any comers. If Intel doesn’t, AMD gets a reprieve.
3. IBM might be capable of shipping 2.4 to 2.6 GHz 970s in Q4 03 as the chip and 130 nm fab matures together. It’s all about increasing the yields.
4. 90 nm 970 chips will probably scale up to 3.6 GHz (after awhile) if IBM can squeeze 2.4 GHz on 130 nm.
5. Last I heard, IBM ramps up the 90 nm fab sometime 1H 04.
6. Dollars to donuts, AMD will ask IBM to fab Athlon 64 on IBM’s 130 and 90 fabs.
7. 2nd-3rd Qtr 2004 comes PPC 980 (400% speed improvement) whole new ball game
8. 2nd-3rd Qtr 2004 PPC970s placed in Apple’s consumer-range lineup.
Joswiak added that in the Intel modifications for the tests, they chose the option that provided higher scores for the Intel machine, …. they would be happy to do the tests on Windows and with hyperthreading enabled, … as it would only make the G5 look better.
Stats really don’t mean crap, but I just assume that people be honest and straight-foward in reporting benchmarks rather than spare our feelings. Anyway, I call B***S*** on this. Does anyone actually think that Apple wouldn’t want to use better stats to their advantage?
“Does anyone actually think that Apple wouldn’t want to use better stats to their advantage?”
Not at the risk of having everyone saying that they cheated.
They went with the option that was most fair and yet everyone STILL said they cheated! The tests were as fair as they could make them.
If they wanted to cheat the tests the performace difference would have been twice as dramatic.
What I find funny about the extreme tech review is that they pulled their entire review out of their ass. Not only did the Spec results outperfom Intels best offering, but the real world tests showed an even greater speed advantage.
Frankly, SPEC means almost nothing. Why do x86 fans always quote them? For one, SPEC is extremely processor-centric – it does not do a very good job measuring the entire system. Two, it is far too easy to manipulate. Don’t you think Intel, AMD, and Dell tweak their compilers? I could go on about what is wrong with SPEC, but I am too bored already. Suffice it to say, it is a test for a processor, not a system.
I guess Apple included SPEC tests since so many people want to see them. Why? Who knows? When I see SPEC, my eyes glaze over (yes, even when they come from Apple). I want to know what my app will do, not some optimized-up-the-yahoo benchmark. As a /.’er said, I don’t run benchmarks much.
I want to know what my purchased-off-the-shelf app on my purchased-off-the-shelf computer is going to do. According to those few tests run by Apple, the dual G5 has a significant advantage. That is what matters to me.
Give me an overall system benchmark – a good one, kinda like a real world app test – and I will start to pay attention. Until then, I want real-world performance estimates.
So to all you PC-whiners that just posted a slew of messages, stop pissing on this board. You’re stinking up the place. Apple has a fast machine, argueably (sic) the fastest (currently).
I find it ironic that he used the phrase: “not indicative of real-world performance”
When, in fact, SPEC is “not indicative of real-world performance”
Application benchmarks, plain and simple. Create a Photoshop action that uses several filters, and time how long it takes to complete. Export the same video clip in Premiere.
I certainly won’t debate that Apple’s benchmarks are incredibly misleading and not in the spirit of SPEC. 3rd party application benchmarks are what I’m waiting for, as they not only show of real-world performance instead of theoretical performance, but also demonstrate how things like AltiVec affect performance.
When, in fact, SPEC is “not indicative of real-world performance”
Indeed, considering ICC is designed specifically to optimize code in the tests used by SPEC CPU2000… similar to nVidia’s cheating on 3DMark
“I certainly won’t debate that Apple’s benchmarks are incredibly misleading and not in the spirit of SPEC”
You are wrong. And I know what you are eluding to… Before you respond I’ll retort by saying that to test the hardware alone, you must normalize the compiler out of the equation — using the same version and similar settings
“you must normalize the compiler out of the equation”
So using an Apple supplied version of GCC vs. a Generiv version for Intel, is normalizing???
How about the fact that they did some register stuff, is that fair?
This is much like the Java vs. .NET petstore (and indeed most benchmarks) – they *overly* optimize the winner, and do a pissjob on the looser – thats how benchmarks work. Unfortunately.
The seemingly endless G5 benchmarking topic is “make work” journalism.
And the myopic focus on small elements of system performance takes away from the fact that Apple has a great new partner for their high-end machines, just introduced the first 64-bit personal computer, and put their first model out in an absolutely incredible case with a very reasonable price tag.
There’s just so much more to look at other than beating SPEC benchmarks into the ground.
Before you respond I’ll retort by saying that to test the hardware alone, you must normalize the compiler out of the equation
This is not possible with a cross-platform benchmark. How does using GCC on both platforms help if it’s been optimized differently on different platforms?
Regardless, the SPEC numbers will always reflect to a certain degree how well the compiler is optimizing code for a different platform. Whether or not Apple (or rather, its subcontractors) sought to intentionally deceive people through its benchmarks is another question, but the point remains that the benchmarks are not indictative of real-world performance.
Surely “Anonymous” you can not argue against the need for application benchmarks… and that’s simply what I’m waiting for as a true measure of performance
Check out
http://www.soundtech.co.uk/news/emagic.htm
(submitted this story, hopefully will be picked up to counter-balance all the FUD flying around by supposed “objective” PC sites)
“We created a processor-intensive workload of projects containing multiple unique audio tracks; assigned five default reverb plug-ins to each of the audio tracks; and tested each platform to see which could play more plug-ins. For Logic, we used the default settings for the Fat EQ, AutoFilter, Chorus, and Silver Compressor plug-ins. For Cubase, we used an equalizer and the default settings for the StepFilter, Chorus, and Compressor plug-ins.”
“The dual 2GHz Power Mac G5 with Logic Platinum 6.1 can play 115 tracks, compared with a maximum of 35 tracks on the Dell Dimension 8300 and 81 tracks on the Dell Precision 650 each with Cubase SX 1.051”
“More impressively, the 1.6GHz single-processor Power Mac G5 played 50 percent more tracks than the 3GHz Pentium 4-based system.”
Now we’ll probably see the PC fanboys arguing that this test is invalid because Sound Technology was using different, optimized apps on each system. No doubt they will now be demanding Apple stand by their “same compiler on both systems” benchmarking (not). Someone please tell me who is being the fanatic here….
Forget about the benchmarks techies love to tinker with. The real ones are the following :
a) will I have to borrow money from relatives to feed myself after I’ve bought a computer, no matter which brand it is ?
b) will my newly acquired computer last long (in terms of years) and perform well enough so that I can get a good return on my investment ?
The usual benchmarks are only marketing gimmicks necessary to lure new customers. Those among us who see computers as tools know that what matters is the ability to give the boss or the client what they paid us for.
Doom3 will be relased for both systems, G5 and PC. So, where is the problem? You`ll install Doom3 on PC and G5 Mac. You will run a Benchmark and count the FPS, with same quality settings. Then, at least, you have a “real world” comparison of 3d-capabilities.
“Windows XP:
How fast does it upload your information to Microsoft?
Windows XP:
How many innocent people get put in prison for each copy of Microsoft Windows that Microsoft sells?
Windows XP:
How much pirated code does it contain? Why does Microsoft argue so vehemently for closed source?”
You forgot the “:)” and *grin*’s because unless its tounge in cheek then comments like that isnt helping anyone. >:-(
Jobs always says, “Look at it <pause for emphasis> isn’t it beautiful”, “I want to hug and kiss it, and run Toy Story on it”
Well I made up that last part
, but we always get the above comments at keynotes and I crack up laughing.
——————–
Also whether the G5 is faster or not, one thing it does prove, is that the Mac faithful will be able to use their tools or favourite apps on a l33t machine.
I am purely a Windows/Linux user who has walked into a few Apple stores in Australia just to mess around with the machines, if I personally found the 15″ powerbook smooth to use can you imagine how fast the OS would be on this machine!
M-A-R-K-E-T-I-N-G
Firstly, everyone needs to be aware that SPEC as a benchmark is specifically designed to be compiled for each platform and be able to be tweaked in compilation by the manufacturer in order to get the best performance from a given architecture.
In short, Apple needs to produce the best result they can (which they have) and AMD, Intel and others are responsible for trying to get the best out of their own systems. All results are then submitted to SPEC, where everyone can see the results.
Optimised compilers are considered fair, as long as they don’t change the actual operation of the benchmark.
The thing is, Apple did their best to produce good results (This is a good thing), but then they produced poor PC results to compare it to, instead of Intel and AMD’s own best results at SPEC’s site (This is bad).
As a result, they compared their best results with results that were, in places less than half of what the SPEC.org results were for the P4/Xeon and declared themselves winners. This goes against the idea of SPEC, that everything should be optimised as far as possible.
Personally, I do think they intentionally fiddled the numbers for the P4 and Xeon. This is simply because if you talk to anyone who runs benchmarks on a major tech site, you’ll see that when the P4 came out, it got shat upon because the FP performance is very weak. The only thing which fixes that is SSE2, which isn’t too terrible, considering that x86 is heading towards SIMD FP processing for the most part. But old benchmarks on the P4 showed it up very badly. Then when applications were compiled and released with SS2 optimisations later on, it suddenly did really well. In my opinion, this is indicative of some poor design decisions on Intel’s part (which is why I buy AMD, because I can get similar FP performance, but across applications that aren’t SSE2-compiled as well), but that is the nature of the chip.
To turn off SSE2, however, will devastate the performance numbers coming from a P4 or a Xeon. If the company Apple employed to do the benchmarking actually know anything about their field, they would have known that.
As for SPEC being a poor benchmark, it is very processor-intensive; there is no doubt about that, but it is one of the few benchmarks that really can be used across platforms and is pretty equal to actual performance comparisons in the real world. To be honest, you’ll have a hard time comparing the performance of a Power4 and a 486 sitting there typing a document in your favourite word processor. You can’t use graphics benchmarks, because they are almost always videocard/AGP bottlenecked, memory benchmarks are meaningless if the CPU can’t take/give the data in time with it, so-called productivity tests are fairly poor because they often rely too much on the disk subsystem.
So where does that leave us? Application benchmarks and synthetic benchmarks (such as SPEC). Apple are fond of their application benchmarks, specifically Photoshop and other applications that they produce/have paid for optimisations in. Let’s face it, Apple’s tweaking of Photoshop filters is common knowledge, and that’s why the wheel them out every single time. In theory, matrix-based operations on large datasets are a good test of processing power (I’m a PhD student and these are what I do regularly), however, they are easily optimised to different platforms, and SIMD (Single Instruction Multiple Data) instructions such as Intel’s SSE2 are designed for it, but not always used. I don’t know how far Adobe have optimised their filters for SSE2 or the design of the P4, to be honest. I do know that given free reign to optimise a filter as I pleased, I could make a benchmark that said anything I wanted it to. Photoshop is all about interactivity with the user, and that by its nature, requires waiting for a user to perform an action; it was never meant to be blindingly fast, or a benchmark.
My personal belief is the following:
The G5 machines are comparable to current P4s/Xeons and Athlons. Where the Athlon lags the P4 by around 5-10%, I would say Apple’s new machines are maybe in the 10-15% behind range. Fairly close. Close enough.
Let’s admit a few things:
– By and large Mac buyers aren’t buying to do high-end calculations.
– Mac users buy Macs because they want a box which looks pretty (nothing wrong with that, I’ve got a neat little Shuttle box that I spent a bit more on!), they like the Apple approach to the business, they like OSX and they like the little assumptions that the OS makes when they perform tasks. All OSs make assumptions, but some suit people better than others.
– It’s also probably fair to say that a good number of Mac buyers buy so that they can feel it is superior and that they are. It’s life, it’s okay.
– Macs do cost more than PCs. That’s a fact, that is life, economies of scale exist. At the end of the day, though, Mac users simply think “Stuff it, money is money and it’s more worthwhile to me to have a machine I enjoy more”.
– Macs are generally slightly slower than PCs. That’s not going to change, but it doesn’t matter, because you don’t do high-performance tasks on them for the most part. Most PC users don’t, either.
– If you wanted your Mac to run faster, you’d be running OS9 not OSX on a shiny new box and you would run it on newer hardware as well. OSX is a processor/video card/memory hog. Quartz’s idea of having each window as a unique screen-sized buffer in memory is, to be frank, scary, and on my machines, where I have lots of windows open, I wouldn’t like to exchange my memory for folding/shrinking/twisting effects. But Mac users rarely have lots of windows open, so it doesn’t matter.
– I’m a web(designer|coder)/designer/(application|server) coder and I dabble in more than a little artwork, so I get to see the range of performance characteristics. At the end of the day, OSX really isn’t up to the complexity of tasks that I perform (even just handling large numbers of windows can be tricky sometimes, and I do use a lot!). That’s fine; it just means that OSX wasn’t designed with me in mind, and I can live with that, and doubtless, Apple can too. I end up using Windows XP for surfing/admin/games, but I don’t like the assumptions it makes, really; which is why I use Linux for work, and I use the KDE and tweak it heavily until it does things the way I want it to to maximise my productivity.
At the end of the day, buy a machine for what you need, friends. If you want speed, then get a PC; consider Linux as well, because the speed and stability are better than XP/98. If you want a pretty machine, have a bit of money and you enjoy the feel of OSX, get a Mac. If you are short on money, or you prefer to be frugal with your computers, buy a PC.
If your life depends on whether Photoshop does a filter 50 times in 5 seconds less on a Mac than a PC, then you have a crappy life/job which you need to fix. If you wanted it done as fast as possible, you wouldn’t use Photoshop (which still has impressive startup times) which hogs memory like nothing else. Get another application.
In short, PCs are a bit faster, and for most of you, it makes no difference at all, so stop worrying about it. Mac users, admit why you actually buy your machines and be happy with it. PC users, leave them to it and get on with what you need to do.
All I have to say to iWhiners out there is that you people never know when to quit…
Apple always saying that Dual G4/1.42 was faster than P4/3GHz and you never did the noise that currently are doing! Damn! Apple must doing something right, this time around!
Kicking some Dark Side A$$ that is! :p
And I bet all those Photoshop, Mathematica, eMagic, Luxology, Blast, etc real world tests are fake too! Or the Adobe, Luxology, Mathematica, et al guys supporting Jobs hype on stage at the keynote were fake too!
Damn! We are talking about the same Adobe which recently created that PC Preferred fiasco for crying out loud!
Grow up iWhiners! This G5 eats your silicon for breakfast without even breaking sweat and is lower in price too!
Boom!
You took a lot of time to produce a lot of rubbish.
Quaint saying, but meaningful in the context of this whole ‘debate’.
Fact: A G5 is a massive improvement over the G4
Yes. But Apple wasn’t comparing with the G4, but with a Dell Dimension.
Fact: The new bus arch is a massive improvement over the G4
Yes. But then again, g4 wasn’t a big deal in that area in the first place.
Fact: MacOS is the most rapidly evolution of any major OS
Good thing you put the “major”
Fact: 2999 for a dual 2Ghz G5 is competitive with commodoty PC crap
Okay, for PC commodity “crap”, you can get an almost equal system for about $1000 less. But for quality* ones like from Dell, AlienWare, Sony, etc., it is about the same price.
* many times having even more quality than Apple.
Your “facts” are just mere opinions.
Your counter arguments are just mere opinions too..
Except offcourse you bought 100 systems from Dell, Alienware, Sony and Apple, and see which ones break down more quickly.
The PowerMac G5 primarily beats that Dell on bus speeds (from 1GHz bus to serial ata), where the processor power is more or less the same. But because bus speed is alot more important than the processor (look at a Sun Blade for example), a G5 can and will easily beat that dual xeon dell in most applications.
G5 is an improvement over G4, this is fact and they do compare the two in both the keynote and tech papers.
167MHZ G4 bus versus 1GHZ G5 sounds like an improvement.
“* many times having even more quality than Apple. ”
Like plastic case, no computer controlled fan system, yeah PCI-X in a Sony. Good one.
Why is it that WWDC and MacWorld seems to be more popular with PC people than Mac people?
Now that the G5 and specs turn out to be true the lamers and PC Sanford and Sons are running damage control. Watch comments on this news thread grow. The majority of which will be PC users like all the other recent Apple news articles! Its so funny that Apple is getting all this attention from the other camp.
So if anything Apples marketing spiel DOES work and is generating a buzz. People will go into Apple stores to see what this is all about since reading junk on the internet is only half the story.
Get in front of these new G5 and you will see how nice and fast it is unless you are ranjar who is enamored by flimsy plastic cases and cheap silver accents and think Sony PCs are high end and thinks that 1GHZ FSB is boring but 800MHZ FSB is better and doesn’t think that G5 is an improvement over G4.
I still don’t trust apple’s benchmarks. As I stated before, a third party needs to test the performance of a G5 system and a Xeon system by running tests using normal, sotre bought copies of software that is made for both operating systems. Photoshop. Premiere, After Effects, Lightwave, Avid XpressDV, etc.
everyone rigs those benchmarks which is why they are useless. Real world apps that is what we need. Personally, i don’t doubt for a second that the G5 will burn with logic audio, bias DECK, or MOTU digital performer and thats all i care about.
A PC running cakewalk, cubase, or sonic foundry’s vegas and direct X effects really can’t compare.
Jesus, don’t you read? The tests WERE done by a third party, and at WWDC Apple did live head-to-head with Photoshop, Mathematica, Logic vs Cubase, etc. and still won handily. The problem with PC zealots is that they just ignore facts which don’t support their dubious claims. Having been on your side of the fence for the P3 vs G4 debate, I can sympathize–it’s hard to have the inferior line of systems–but that doesn’t change the facts.
I’m not concerned about minute differences, as long as they are reasonably close in performance that’s good enough for me. I personally have a very difficult time telling the difference (performance wise) between my 2.5GHz P4 work machine and my 1900+ AMD machine at home. But then again, I’m running OSX on a G3 500 12″ ibook and I like it
The SPEC part of it. All these critics are ignoring the app part. Go watch the video. Dual 970 crushed dual xeon over and over on photoshop, mathematica, etc. It was embarrassing for the xeon, frankly.
On http://www.apple.com/powermac/performance the benchmark hyped was SPEC – not Photoshop tests, but SPEC. It is right on top the page, for some apparent reason, and the fact that Apple may have optimized their hardware for performances, but do otherwise for the Dell speaks volumes. More proof? The main page – http://www.apple.com/powermac/ – they didn’t show Photoshop benchmarks. They didn’t show Cubase benchmarks. They didn’t show Blast benchmarks. They didn’t show HMMer benchmarks. They showed SPEC. That’s a big half if you ask me.
And in the video, there was more said about the fabrication process then about Photoshop by that Adobe executive. The funny thing is that the PC used for the benchmarks is that the whitepaper never said anything about HyperThreading being enabled – which have some performance gains, only the ancient SSE2. In other words, Apple left out a selling feature that Adobe has yet to optimize for in the whole benchmarks.
Why is it that when ever I say anything I am instantly called a PC zealot?
I’m not. I never said that the G5 is slower than a Xeon system. I never said the the new powermacs sucked. I never said ANYTHING anti-apple.
All I said was that you can’t judge by the benchmarks that apple is showing you. I wouldn’t trust Intel or AMD’s benchmarks either.
You need something done by a thrid party (who is not paid/requested to do the benchmarks by apple or intel). With the normal versions of the software and a normal system just like I would get it I went and bought the systems right now.
Rajan: “Okay, for PC commodity “crap”, you can get an almost equal system for about $1000 less. But for quality* ones like from Dell, AlienWare, Sony, etc., it is about the same price.”
yea, almost equal except runs real world apps like Photoshop, Mathematica, at half the speed.
Assuming that Adobe, Wolfram, etc. are in on some massive fraud on those real app tests and that their VERY POSITIVE comments on the G5 performance at they keynote were just made up is wishful thinking at best, delusion at worst.
AMD doesn’t make any compilers. They should make compilers that have some decent optimization for their processor, but alas, they still do not have any compiler.
What a bunch of crybabies!
Just get over it you silly wonkers, in 6 months a new x86 chip will come out and we will talk about this again, and then in 12 months we will see another new G5 chip come out and we will talk about this again, and so on and so forth till the end of time (or the apocalypse anyways)…. jeez.
Personally I think it was a fair test because an INDEPENDENT firm did the testing
And I bet all those Photoshop, Mathematica, eMagic, Luxology, Blast, etc real world tests are fake too! Or the Adobe, Luxology, Mathematica, et al guys supporting Jobs hype on stage at the keynote were fake too!
Considering the close partnership between the companies and Apple (and IIRC, Apple also bought Cubase, for good measure), and for many of these companies the Mac market is quite wide and profitable – it isn’t a big scretch of the imagination.
Damn! We are talking about the same Adobe which recently created that PC Preferred fiasco for crying out loud!
For a product who had its market severely cut up on the Mac because of Final Cut Pro. In other words for Adobe, it was a little payback kinda deal. But Photoshop-wise, Adobe have been more Mac-centric than PC. Consider the fact that Photoshop was available on Mac OS X faster than it was available on Windows NT 3.5 (if there was a version for that Windows NT in the first place). Consider the fact that Adobe took a relatively short time to introduce some optimizing efforts for Altivec when G4 was first release, but was rather relunctant for SSE2.
Shouldn’t the headline be, “Apple putting out a great computer makes the people who want one but can’t afford it jealous?” Funny how everyone gets their back in the air about the “comspiracy” behind a computer company actual painting their performance results in a positive light – that is – whether or not you actually think they’re accurate. Which, on their own merits, I’d say of course they are. I don’t know, if I had a computer company and was going to lie about perfomance results, I would come up with more outrageous results than that.
ALAIK the 64-bit version of GCC for x86-64 is made jointly by AMD and Linux guys. More Information is available at http://www.x86-64.org.
I think the x86-64 architecture should be a pleasant platform for everyday use…
It is REALLY important to some people to know that they have the fastest computer, fastest car, pretiest girlfriend, and the biggest penis. And when something comes out to make them feel second best in ANY of the above categories, they will go to any length to try to regain their throne. Actually, their whole life would be funny if it weren’t so sad.
InKonu
Your counter arguments are just mere opinions too..
No, they weren’t. Take for example my first counter argument – Go to http://www.apple.com/powermac/ – the first benchmark they show is between the top end, a Xeon, a P4. Then take a look at my 2n counter argument – how fast was the G4’s bus speed in comparison with any of its x86 competitors? Then my third counter argument – you to any clone PC shop of your liking, give the specs of that Dell, you would definately come out with a price quotation significantly lower than of Dell. Remember, these clone shops own about 55% of the global market.
Except offcourse you bought 100 systems from Dell, Alienware, Sony and Apple, and see which ones break down more quickly.
No, I don’t have any of their systems. But I know people who got them and have no complain about their quality (they may complain about how they got rip-off because I got a better machine for a lower price from the average clone sho[, but no complains for quality). And in the case of Dell, I actually heard a lot of recommendations merely based on their services.
How many Macs have you owned in relative to the PCs you have owned to make such a judgement yourself?
The PowerMac G5 primarily beats that Dell on bus speeds (from 1GHz bus to serial ata), where the processor power is more or less the same. But because bus speed is alot more important than the processor (look at a Sun Blade for example)
When Pentium 4 first came out, it had a higher clock speed, and a higher bus speed than Pentium 3, and to the larger extend, Athlons. Was it faster? Nope, in both cases. They may play a role, but their speeds don’t play a definate role.
And a Sun Blade has a whole lot more than just a fast bus.
a G5 can and will easily beat that dual xeon dell in most applications.
Maybe they could. But I’m not one to trust Apple for such benchmarks, after they using the worst combination for the PC in their most hyped benchmarks, besides the fact that they also lied previously on applications benchmarks (only the SPEC benchmarks is done by a “independent” company, the rest are, to quote, “Test performed by Apple in June 2003 using preproduction Power Mac G5 units and, with the exception of HMMer, application software optimized for the Power Mac G5)
And besides, G4, when it came out, was rather much faster than x86. And after some short months, x86 overtook and manage their position until Sunday.
But if you really look what’s going on the ground, companies like SuSE are really just doing much of the work. By comparison, Intel wrote their compiler from scratch, and have put a whole lot of money in it. AMD did little more than provide documentational help to the developers.
Apple routinely distorts their benchmarks. Besides, having outrageous benchmarks results that have absolutely no hope in being replicated in real life does a whole lot more bad for your company than it would do good. However creating a loopsided benchmark is really a much much better solution, no? And no litigation problem too, because technically, you didn’t lie.
So just because everybody else cheats on benchmarks it makes it okay for Apple to do it?
I’m not saying Intel, or AMD, or Sun, or Dell, or HP, or <insert lying computer company> are any better or worse than Apple, but it’s ridiculous for them to have to lie and cripple the opponent in a benchmark. Just because it’s done by a third party doesn’t mean the third party didn’t cheat for them.
<sarcasm>
This third party wouldn’t benefit from giving Apple a heads up in any way, would they? nahhhhhhhhhhh, couldn’t be.
Future business with Apple being at stake, not to mention the benchmarks that will be done by chip makers on the PC-side trying to “beat” Apple
</sarcasm>
I’m not saying that Apple’s any worse than the others, but they’re also not right to do this.
And I’m sure these were probably the best of a series of benchmarks, once again, not that other companies don’t do the same, but it’s still low.
My beef is with everybody who takes Steve Jobs’ words as the new gospel. Steve is not Jesus, okay Mac people, he’s mortal, he’s human, and at the end of the day, all he cares about is the all-mighty dollar
LOL, that is the best post in this entire thread!!! Everybody needs to read that.
Everyone distorts benchmarks. And, like someone said, what is Jobs supposed to do up there, say the Mac can’t cut it??? We’ve got to move on.
The important thing is how OS X and applications run on the G5. That’s the only thing that matters – everything else is trivial.
The only other thing that matters is that the G5 has room to grow (Jobs saying 3.0 in a year). This is great news for computing. Some of you are the most mirthless and joyless people I’ve ever seen, going round and round and round about the same trivialities.
I’m sure the test were accurate. This is by far the best performing Mac in history and the 95 hates it. Well they hate everything Apple puts out, period!
Duu the 970 is based on the Power4 PPC, remember. IMB’s own Specs(1.8) said it was fast as a 3 gig pentium. We are talking about a Dual 2gig 970 that is optimized for dual processing. Sorry Losers!
A major magazine will get their hands on these macs and do a review and test. Basically your going to eat own words in the end!
Thats good news for Linux users that Apple’s new hardware ran Linux software faster then the Dell. It would be a great system to dual boot into whatever flavor of Linux you want. Heck it would probably run Shake faster too.
Wonder whatelse it will run faster
it is easy.
just buy the damn thing.
when there is something you want you justify it when you buy it.
if you want a mac, just save up for it or finance it.
I have a realy hard time that so many people on OSnews are actualy doing very well debt wise, they may be.
when you realy realy want something, every few years youy can justify the price, and the good thing about the mac is that tehy last about 2 years longer than PCs on average…..I have a 3 or 4 year old G4 400 (it was my brothers who gave it to me when he bought his powerbook so I don’t know exactly how old it is) and it has never been upgraded but runs 10.2 as well as my duron 1.2 and my 1.1 cel laptop.
bytes256: “My beef is with everybody who takes Steve Jobs’ words as the new gospel. Steve is not Jesus, okay Mac people, he’s mortal, he’s human, and at the end of the day, all he cares about is the all-mighty dollar”
Actually it’s more than money that motivates these people since they already have more than enough. It’s also not the case that Steve has nothing else to do, since he also has Pixar. Steve is probably driven by ego, and the desire to beat MS. He thinks he can do it by making better products, at least better in the sense they are more accessible and usable. He might be wrong but that’s what he’s up to.
> If your life depends on whether Photoshop does a filter
You haven’t been in many Graphics departments, have you?
> SPEC
Except sizeof(int) on Intel POS is 4 and sizeof(int) on 970 is 8. So, even may the SPEC numbers be lower, the G5 is pushing around twice the data, so the G5 is still faster…!
Actually it’s more than money that motivates these people since they already have more than enough. It’s also not the case that Steve has nothing else to do, since he also has Pixar. Steve is probably driven by ego, and the desire to beat MS. He thinks he can do it by making better products, at least better in the sense they are more accessible and usable. He might be wrong but that’s what he’s up to.
The same thing could be said about Bill Gates or Larry Ellison or Scott McNeally.
My point is that Steve is an excellent speaker/motivator/salesman but he’s not selling Macs just for the betterment of the world as some Mac zealots seem to think.
“My point is that Steve is an excellent speaker/motivator/salesman but he’s not selling Macs just for the betterment of the world as some Mac zealots seem to think.”
Just as this article points out, Mac G5 runs Linux programs faster then the Dell. So he probably will be getting a lot of Linux switchers again.
The stats were not distorted a little, they were not distorted a lot they were completed in the most fair way possible.
Apple chose the GCC comiler not because it performed better with the GCC compiler but because the compiler was available for both platforms. Yes, Intel would have got better stats if they had used a different compiler and yes SO WOULD THE G5!
What I fear most is that whatever suposedly independant reviewer will compare Apples hardware with the GCC compiler against x86 with their preferred compiler. (IE, not offer the same cortesy of fastest the best compiler on the G5).
This new test will end up giving the x86 machine the preferential treatment and THEN all the PC fanboys will say, “HAH! SEE, APPLE MUST HAVE CHEATED BECAUSE THE INTEL SPECS WERE INDEED FASTER!!!! LEWSERS!!!!!!”
Because this spec will be the only one where the x86 came close (although probably not necesserally beat the G5) we’ll end up simply rolling our eyes.
1) Marketing is lying. You convince someone that he needs to buy something he really didn’t need. So benchmarketing lying is part of the game.
2) ANY benchmark is FLAWED. Period.
3) SPEC, as someone said before, is the best people got for benchmarks – kind of objective *and* cross-platform. I personally don’t like SPEC’s policy to accept results got by optimized tests, but…
4) For me, “I only care about application benchmarks” sounds like “My benchmark favours the machine I like, and f* the rest”. So objective, eh?
5) Nobody had access to the Dell Dimension’s specs. It was an ordinary one? It was optimized? It was de-optimized? Nobody knows. OTOH the G5 was optimized.
5) I would have said something about Apple zealots, but as ferocious as they are about it, I prefer my physical integrity.
Flames go to void().
During the 60s Italian supercar manufacturers routinely fitted far more powerful race spec engines to cars being tested by motoring magazines. In addition Ferrari and Maserati would publish wildly exaggerated internal performance tests. This had several effects firstly Ferrari and others got publicity for their “160+mph” vehicles (the actual model sold to customers would often be 20 mph slower). The foreign exotica owners would be really pissed off when they found out that their ultra expensive machine was sometimes slower than a vastly cheaper American muscle car. What did the customers do? They normally vehemently denied that their Ferrari was slower than a warmed over 427 Corvette.
Porsche was different. It always understated the performance of its machines. The customer was happy because their car was faster than they expected and Porsche got a reputation for honesty.
Either the Apple benchmarks are true and Apple customers are getting very fast, stylish and somewhat expensive machines or the benchmarks are false and the Apple machines are simply non-competitive.
The G5 won’t be out for 2-3 months. The Athlon64 may beat it to the market,. The price of Intel and AMD CPUs will be less and newer and faster x86 hardware will be on the market.
Let’s see what happens.
The stats were not distorted a little, they were not distorted a lot they were completed in the most fair way possible.
Hm… You Apple zealots are SO funny…
So I am a car maker and I want to hype it, I make some undisclosed tests (with my optimized car vs eg a factory-default Ferrari) and I show the final resultas, saying “my XPTO car is better than any Ferrari”. No car magazine will buy my benchmarketing.
Substitute “I” for “Apple”, “my car” for “G5” and “Ferrari” for “P4”. The results will be the same.
To be fair, I’m starting to being sick of the computing industry.
>> In short, Apple needs to produce the best result they can (which they have)
That is not true! GCC does NOT compile the fastest code on PPC. In fact, CodeWarrior and IBM übercompiler are both much faster.
It would be interesting to see a benchmark of ICC on the Xeons vs IBM übercompiler on the G5s.
Also, for those asking about the specs of the Intel machines, download the pdfs off the veritest site, they have some specs of those. For example, both Intel machines were equipped with 2GB of memory, whereas the G5 only had 1.5GB for all tests.
Hm… You PC fanboys are SO funny…
you’ll say anything, be they car analogys or whatnot to distort the issue.
The fact that we guys are stretching so hard to find a flaw ANY FLAW with this independent research study is hillarious. You then tout real world performace over Spec tests not knowing that the G5 beat x86 even more.
We told you this day woould come. We’ve led the speed race the entire time. We got hung up by Motorolla why wouldn’t push PPC to its limits, you claimed it was due to ineficiencies of the articture, you told you you were wrong, and in fact… You ARE/WERE wrong.
G5 outperforms the best x86 has to offer in both Spec AND real world tests
“It would be interesting to see a benchmark of ICC on the Xeons vs IBM übercompiler on the G5s.”
It would be interesting… however inconclusive, as that would only show the efficiency of the specific compiler and not the efficiency of the processor. I would expect the tests to end up very similar. To create a balanced test it is impeative that both computers use the same compiler.
How much do you want to bet that the PC fanboys, when they get a hold of this hardware will compare the two using Intel’s compiler on x86 and the GCC compiler for G5 and that THIS benchmark will supposedly demonstrate the Apple lied. I’ll bet anyone $500…. I’m series oabout this.
…exactly and real world test the G5 out performs. There are a lot of unverified test on the G5 prototype and it backs up what we saw on monday. I care in real world performance anyways. I’m a Photoshop and Final Cut user and Apple has made a beautiful machine for me to work on.
PC Zealots, Risc is good!
We all know this can’t and won’t last, if it’s even true now.
Remember, G5s aren’t on the market yet, and Intel and AMD will release at least another chip each by the time it is released.
Let’s all wait til the end of the year and see who the hell is ahead.
(and IIRC, Apple also bought Cubase, for good measure)
Apple bought eMagic (Logic), not Cubase. Steinberg (Cubase) is owned by Pinnacle.
Just an FYI…
Hey!! Stop arguing you shits! Now why won’t the link to g4 to g5 comparisons work?! I’m seriously considering getting a dual g4 now, any help?
Has anyone actually read some of the arguments that have been made that Apple is cheating? They are so thoroughly ridiculous its unbelievable. Apple has gone out of its way to make this as fair as possible, to the point of even favoring the Intel and Xeon in their tests.
Myths:
– GCC is faster on PCC than on x86
FACT: GCC has existed on x86 a lot longer than it has for the PPC, and has been optimized longer for the x86 as well. Both processors have better compilers available as well, however Apple didn’t benchmark compilers with this test, they benchmarked processors. There real world tests benchmark compilers, AltiVec, system controllers, etc.
Greg Joswiak, VP of hardware marketing at Apple “said Veritest used gcc for both platforms, instead of Intel’s compiler, simply because the benchmarks measure two things at the same time: compiler, and hardware. To test the hardware alone, you must normalize the compiler out of the equation — using the same version and similar settings — and, if anything, Joswiak said, gcc has been available on the Intel platform for a lot longer and is more optimized for Intel than for PowerPC.”
– Apple disabled SSE2 to cripple the Pentium and Xeon
FACT: Apple used the flag mfpmath=sse, which does in fact enable SSE2 on the Pentium 4 and Xeon. People are just assuming otherwise without knowing any better.
– Apple disabled Hyper-Threading to cripple the Pentium and Xeon
FACT: Dell disables Hyper-Threading for the same benchmarks that Apple did, because you get higher scores with it disabled.
http://www.dell.com/us/en/esg/topics/power_ps3q02-khalid.htm
http://www.specbench.org/cpu2000/results/res2003q2/cpu2000-20030404…
– Apple used Red Hat Linux 9.0 instead of Windows XP to make the Pentium and Xeon look slower.
FACT: Apple chose the tests that gave the *higher* scores to the Pentium and Xeon. They scored higher using Red Hat Linux 9.0 instead of Windows XP.
http://apple.slashdot.org/apple/03/06/24/2154256.shtml?tid=126&tid=…
“Joswiak added that in the Intel modifications for the tests, they chose the option that provided higher scores for the Intel machine, not lower. The scores were higher under Linux than under Windows, and in the rate test, the scores were higher with hyperthreading disabled than enabled. He also said they would be happy to do the tests on Windows and with hyperthreading enabled, if people wanted it, as it would only make the G5 look better.”
More myths dispelled at that slashdot link above. If you’re really interested in how the PowerMac G5 stacks up against a PC, and not just the PowerPC 970, take a look at:
http://www.apple.com/powermac/performance/
http://www.apple.com/powermac/graphics.html
The dual 2GHz PowerMac G5 blows away the $1000 more dual 3.06GHz Xeon Dell on real world tests in Photoshop, Mathematica, music editing and playback, BLAST DNA sequence matching, HMMer Genome sequence matching, and Quake III Arena 1.32.
then look at these, where the reference Mac is a 1 MHz Mac and the G5 is compared to it. It’s Mac vs. Mac:
http://www.thinksecret.com
Do you people honestly have nothing better to do with your time?
“Do you people honestly have nothing better to do with your time?”
With all the time these computers save us… we have more time to counter the trollings on this board. ; P
http://www.theregister.co.uk/content/39/31416.html
Does any one know how do both machines compare in power usage? Dell vs G5? This is something that is getting quite
bad these days. I think IBM and Intel should spend more recources on cuting down power usage. Well I know that they are dooing it but at the same time increasing GHz. Which effectivly increases overall power usage. Just thinking about it 10 years from now not only state of California will have power problems. Also more heat produced by these machines more air conditioner needs to work. Over all more power to use.
“To test the hardware alone, you must normalize the compiler out of the equation”
“In the G5 modifications, they were made because shipping systems will have those options available.”
So what are they testing here, the hardware or the shipping systems? If they were testing the hardware, they should have tested the G5 with Yellow Dog Linux and no special libraries. If they were testing shipping systems, they should have tested the Dells running Windows using either the Intel or Microsoft compiler (gcc is rarely used for shipping Windows apps).
I am amazed by Apple fans’ lack of skepticism. Too many swallow what Jobs says hook, line, and sinker. This is why these misleading benchmarks piss me off. For some reason the Apple fanboys actually believe them. Do you have any idea how many posts I’ve seen claiming that a 1GHz G4 is faster than a 3GHz P4?
The sad part is that this controversy has very little to do with how good the G5 is. It is about honesty. Apple has a great processor with the G4, but their benchmarking definitely looks suspect. The truth will eventually come out when truly independent tests (not paid for by Apple OR Dell/Intel) come out.
I cant find the link, but I know that the G5 is significantly more efficient (IE less power) than the G4 which was already very efficient with regard to power consumption. In comparisons to P4s the G4 used significantly less heat. So the use of the G5 will increase that ammount exponentially.
I’ll continue to search for a link. If sombody beats me to it, go ahead and post it.
“So what are they testing here, the hardware or the shipping systems? If they were testing the hardware, they should have tested the G5 with Yellow Dog Linux and no special libraries. If they were testing shipping systems, they should have tested the Dells running Windows using either the Intel or Microsoft compiler (gcc is rarely used for shipping Windows apps).”
I suppose a better test might have been to test both hardware platforms using the same OS and the same compiler, but your assumption is that somehow the gcc compiler was optomized for PPC and not x86. GCC has been around a lot longer on x86 than it has on PPC and if anything it has had more refinements done to it than a PPC version. Lest you forget, Joswiak conceded that the Dell numbers would be higher with the Intel compiler, but that the Apple numbers could be higher with a different compiler too.
“I am amazed by Apple fans’ lack of skepticism. Too many swallow what Jobs says hook, line, and sinker.”
If anything, you should be amazed at the skepticism on the PC side. Far too many individuals were touting an unfair outcome before they had any facts. Most went so far as to tout INCORRECT facts. Now that the dust has settled and the trolls went under their bridge it is becoming ever more apparent that the tests were in fact done in a fair manor.
“This is why these misleading benchmarks piss me off.”
I guess its statements like these that piss me off, as the statement wasn’t misleading in any way shape or form.
“For some reason the Apple fanboys actually believe them.”
For some reason the PC fanboys actually don’t believe them
“Do you have any idea how many posts I’ve seen claiming that a 1GHz G4 is faster than a 3GHz P4?”
For some tasks it is. Not most, but certinly many.
“The sad part is that this controversy has very little to do with how good the G5 is. It is about honesty.”
Exactly. Thankfully the parties involved in the benchmarking process were very honest.
“Apple has a great processor with the G4, but their benchmarking definitely looks suspect.”
You are correct that the benchmarking with the G4 was misleading (although never false). Thankfully the G5 settles any concerns that any might have felt about PPC and its future.
“The truth will eventually come out when truly independent tests (not paid for by Apple OR Dell/Intel) come out.”
I fear that these tests will use GCC compilers for Apple hardware and Intel compilers for x86. When the results come back with x86 benchmarks which differ from those that independant research group Veritest came to, all the PC fanboys will cry foul, when in actuality, GCC compilers were used to giver parity to both platforms.
Remember, Greg Joswiak conceded readily that the Dell numbers would be higher with the Intel compiler, but that the Apple numbers could be higher with a different compiler too.
The bechmarks ARE acurate.
“Oh well, I can’t currently, and probably never will be able to justify the expense of a Macintosh. It just isn’t worth it when I can get free OS’s and fast, cheap hardware.”
You obviously are not a creative person.
If you used your computer to make your living creating media of any kind, you would see the Mac as the best, most cost effective tool for the job. Why do you think it is used 10 to 1 in the entertainment industry?
This is where the average Joe gets lost in comparing the two platforms. One one side you have a commodity based mind set (lowest price quantity over quality etc. etc.) and on the other you have the creative professional (Real world speed, ease of use, cost of ownership, return on investment etc.etc. etc.)
It goes without saying that you will rarely see one type hanging out with the other since they are polar opposites when it comes to their view on life and on the value of a computer as a tool.
What makes anyone think a Windows drone is going to ever accept the fact that an Apple computer is better than anything they can buy in the Wintel camp? To them, the advantages are just as non-tangible as their creative ability.
As long as they can run Doom III faster than their neighbor they are happy campers. ;o)
It takes alot more than processor speed to run benchmarks. For one they arenot telling anyone the specs of the Dell models used, no RAM no bus speed nothing and I asked and I got ignored and when I asked again, I got a look. If Apple has nothing to hide tell me the specifications of the Dell machines, they wont. Second there was alot of misreporting at WWDC. They claimed that they are the only ones that have backwards compatibility with 32bit, that on the x86 side you need to run a 64bit OS, or run the OS and Apps in an emulation mode, an untruth because the AMD Opteron offers the same functionality as the Power4 that Apple is using. Overall, they are really nice machines they are speedy, a definate enhancement in Apples line. Also, the Xeon is dead Intel has abandoned it like Moto abandoned the PowerPC development, what I want to see is a test against Itanium and AMD Opteron, to me those are the benchmarks that matter because that is what people are moving too, Xeons pretty much dead as is thePentium 4. And I know, its an Apple sponsored event and Apple is going to evangalize and lie and cheat and steal in order to push their product. Im well aware of that and have gotten used to it. The main thing for me is the price. And realistically what am I going to do with a 64 bit Mac. These machines are mainly focused on the highend graphics worktations and not the desktop users, how many working class slobs do you know that are going to pump out the 3 grand for a G5, I do not know many that can and that will. I must say tho, I am impressed with Apples offings and I have always been their harshest critic and if they can impress me then I think that they have a winner. I just wish they were more open on the Technical side. To be so hush hush about test machines and technical specifications to me makes it look like Apple has something to hide and I think thats where all this mess started.
I suppose a better test might have been to test both hardware platforms using the same OS and the same compiler, but your assumption is that somehow the gcc compiler was optomized for PPC and not x86. GCC has been around a lot longer on x86 than it has on PPC and if anything it has had more refinements done to it than a PPC version. Lest you forget, Joswiak conceded that the Dell numbers would be higher with the Intel compiler, but that the Apple numbers could be higher with a different compiler too.
The OS isn’t even supposed to matter for the benchmark (although it could in some ways, it’s not supposed to). Fortunately, there are already plenty of published benchmark results to look at, and if anything this test shows that gcc is absolutely not the compiler to use for this benchmark on x86. On the PPC side, gcc’s PPC development over the last few years has primarily come from Apple (and a bit from IBM as well), who supposedly uses the compiler to build OS X. If there is a better compiler to show SPEC results on these processors, then we should all be assured that IBM will be publishing such results, as they have with their previous CPUs on their compilers.
“I am amazed by Apple fans’ lack of skepticism. Too many swallow what Jobs says hook, line, and sinker.”
If anything, you should be amazed at the skepticism on the PC side. Far too many individuals were touting an unfair outcome before they had any facts. Most went so far as to tout INCORRECT facts. Now that the dust has settled and the trolls went under their bridge it is becoming ever more apparent that the tests were in fact done in a fair manor.
The easiest way to determine that the tests were performed in a manner that was at least slightly skewed was to look at the existing SPEC numbers for the x86 processors mentioned. Apple’s benchmarks will always be questioned by people using x86 processors, but this is really the first time in recent years that they’ve even bothered to explain themselves in the least, and part of that is because they have to disclose a certain amount of information in order to even publish SPEC results. There’s a reason that Jobs and Apple are known for the reality distortion field, and the benchmarks Apple releases are a good part of that.
Apple makes good products, I have no doubt about that (and I was partially raised on their products), but their marketing is worthy of as much attention as the ‘news’ rags in the supermarket checkout line.