Apple's Benchmarks misleading? - Page 5

Agreed.

Let me also add that this is only the *beginning* for this new line or processor from IBM. I'm sure we'll see additional compiler optimizations and applications developers willing to take advantage of the CPU's features.

I also remember Steve Jobs mentioning that the 970's were built for massive parallelism in mind -- full SMP support. I then remember him mentioning that the powermacs have 2 processors and that it's only the beginning ... or something to that effect. Perhaps someone here can recall what Jobs said about SMP. I may have gotten it wrong.

As far as the Opterons go... I just don't see them hitting the CompUSAs and CircuitCity type outlets any time soon, but considering how much trouble AMD has been having with getting their tech to market, it does make you wonder just how much longer they can keep pushing the envelope.. the same goes for Intel. This CPU war has caused them to push harder and faster in a race to bring the fastest CPUs to market as quickly as possible no matter what the cost. How much further does either company have with respect to pushing the envelope? Remember, Intel was working on the Itanium for a VERY long time. I suspect that they saw the end of the x86 line coming sooner than they would have hoped -- that's probably why they started the project. Now AMD goes and brings out AMD-X86-64 -- x86-64 is something that Intel *didn't* want to have to deal with. I can't imagine how it will pan out for the Wintelon crowd. I don't think they've been thinking about those types of problems.

--
Ed

Apple have yet to play their 'Panther' ace.

And by the time they do a bumped 970 to 2.5 gig should be imminent.

As for the Opteron's Lightwave performance...I'm left distinctly unimpressed. Out flanked by the dual 2.8 gig Xeon. I can see the G5 doing well on Lightwave benches. I can't wait for Newtek to optimise for the G5!

In fact, in the applications you'd expect Apple doing well in...the Opteron seems very mediocre. Not the monster that the guy at that SmugVideoTM sight seems to think. If that's next gen'...Apple shouldn't be too worried. And AMD a partner. I doubt Apple would slag their next gen' off when Intel of the Wintel alliance is the real target of the 'bake-offs'. Heck, Apple ARE using Hypertransport after all...

It's all right Opterons/Xeons doing brilliant in a theoretical...or clinical test (WHICH, I may add, has been biased and optimised for PCs for years but you DON'T hear PC whiners or SMUG VIDEO TM or Bi-ASSED Wintel sites mentioned that...)...but Spec doesn't really mean anything to me.

G5 crunches Xeon, dual for dual, in PS test? THAT I DO UNDERSTAND!

Not only PS. Now they have a machine that crunches Xeons in 3D, Mathematica, Motion capture, Audio...s'funny. Sounds like Apple's target market, noh?

Lemon Bon Bon

I agree. I don't see the AMD 64 in consumer towers in volume anytime soon.

This cpu war has taken its toll on AMD. They've had problems keeping up since they beat Intel to the 1 gig mark.

Intel are like a pig in labour with the Itannic.

The 64 bit picture looks like a mess with Wintel politics flying.

(If I was Apple...I WOULD develop a version of 'X' for Apple branded Intel...to drive a wedge between them and M$. As long as PPC is on its new tradjectory, it doesn't matter. PPC is back. It's faster. And Apple's target markets will buy it. Don't see it happening soon as in next year though...)

Intel shouldn't be underestimated, though. IF a 0.09 Prescott comes flying out the door, we may see them hitting 4 gig by the time Apple hits 2.5. (which should still crush any Prescott...hmmmm, can Prescott go dual? Or will it be Xeon only?)

For me, this is the time for Apple to really put the boot into the Wintel crowd. To me, they NEED to move to the 0.09 970s as quickly as possible. Put 0.09 versions of the current speed grades into the current G4 'consumer' tower and iMac and eMacs/cheap headless consumer Mac.

Either way, I'm less worried about Apple than I have been in years. The 'signs' are looking good. Almost all the pieces are in place...

I still don't see the 'market share' builder. Oops. Wrong thread...

Lemon Bon Bon

The monster bandwidth of the G5's Elastic bus pretty much obviates the need for L3 cache. Eventually a L3 cache will be needed, if I understand what some tech gurus have written, but not until the G5's GHz get way up there.

Face it--the Powermac G5 is the POWERmac we've all been waiting for. It's the savior of the Mac platform. It's the "one". VERY cool that Apple has named the PPC 970 "Neo". I hope they keep using that name, it's way cooler than "G5".

According to the Intel timeline Prescott will start at 3.4 ghz and won't be available until 2004 minimum. IBM is already well on the way with the 980 and 970 should ramp nicely as the process improves.

Hold on kiddies...

Nick

So does anyone have any idea based on what we've seen so far as to what we can expect for performance in the video end of the spectrum that SJ so dearly loves? (iMovie, FCE/P, Shake, Maya, etc.)

After how well they were to cut video encoding time down with the G4, what will the G5 herald. Will this be the "system" that gets Apple into the big boys houses like Pixar, ILM, Weta...

Yes i wonder how the dual G5 perform for encoding MP2. I think that the VMX engine of IBM will work great. Hannibal in Ars made only one update of his article about the PPC 970 : he said that the vmx engine of the G5 has less restrictions than he thought. Therefore, due to the superior bandwitch, i expect really good performance in that aera.

No, you won't see Opterons in CompUSA any time soon, and that is why Apple is the first with a desktop 64 bit machine.

The sad truth of the matter is that on the Windows side of the world, 64 bit computing is viewed as something that no consumers would ever want to do. 64 bit computing is viewed as something of a server thing, fir for databases and scientific worksations. It is rather sad because this stifles development work because we all have to assume that you can't have more than 4GB of RAM. This is all Microsoft's fault. MS needs to sell 64 bit windows for the same price as regular windows.

Intel also has some blame to face- Itanium is so ridiculously expensive that it is clear that Intel does not want 64bit computing to get into the mainstream. Only AMD has a sensible approach and they too are charging top dollar in an attempt to recoup their R&D costs. Windows 64 bit is a real mess folks. I greatly appreciate Apple's approach and I hope that it goes well so that MS will get their act together.


Apple will move to the 0.09 G5 as soon as they can. Some things are very difficult to speed up. Apple needs to have iMacs running with 1.8 GHz G5's in a year (they really do!). Apple needs to make 64 bit computing pervasive.

http://www.macobserver.com/article/2003/06/30.5.shtml

Nice one, Brad Peebler. Coming out to bat for the G5. And his Lightwave boys prefer 'X'...

Wonder why that is...?

Lemon Bon Bon

PS. And if Braddy Babes is saying that...I can't wait until his Lightwave benches are out!!!

And advertise that fact. Play the Wintel numbers game.

Apple NEED to get that G5, somehow, into an iMac2 within 12 months! G5 1.8 gig. That's a 3.6 gig Pentium 4 and on THAT bandwidth should hang with any consumer Prescott chip that Intel will have by that time.

If Powerbooks get the G5 by next May...I hope iMac2s get one also. It should make the iMac a kick-ass consumer AIO.

Apple should be bustin' a gut to get their whole line 970.

0.09 should help them get there.

Lemon Bon Bon

Wow! It's truely the end of the dark age when JYD posts something positive about a powermac that I agree with. The fact that all those PC web sites are either silent about it or try to find a reason to dismiss it must mean Apple's got the real deal on their hands. They're worried. I find that ( after 4 years of apologies about tha G4's short commings ) nothing short of hilarious!

Wow LBB.. you beat me to it. I read that story before it even hit the net. At about 5:30 AM this morning EST, my friend Bryan over at the TheMacObserver shot me the link to the article, but asked me not to pass a link around until after it was posted. I had to run to work so it just had to wait ;-) Oh well. Anyway, I'm just glad that these companies are going to bat for Apple. I think it's time to call SPEC (as a benchmark for over all system performance) into question. Perhaps they should retire it?

--
Ed

Unbelievable ... the entire P.C. world has spent YEARS marketing their products sole on the megaherz myth and meaningless white papers, and are now jumping to arms with respect to real world, unbiased comparisons.

tsk tsk tsk

It isn't as if SPEC is entirely wortheless. It is just that it is not as useful at comparing two very different machines. You can compare the SPEC numbers for a G4 and G5 and get rather meaningful results. G5 to P4? That isn't quite as easy.

Steve didn't say first 64bit computer available at compusa or any other retail store. He said first 64bit PC.

So what consumer apps sold in compusa would use >4GB if you had it? The only one I can think of is video editing, which is still sometimes processor bound. You don't have to assume what windows supports. 32bit windows can go as high as 32GB. Maybe if WinXP Home had >4GB support, 2GB DIMMs would be a little cheaper. Even so, a fast computer with say 6GB of RAM costs way too much to be called a consumer desktop.

64bit support is nice but has zero benefit for consumers at the moment.

Photoshop. Video editing. Ripping lots of CD's. People think that these machines are overkill, but they are PERFECTLY targed at Apple's core market which consists of people for whom you can not have enought CPU power or memory. These machines are poorly targeted at most PC users because they don't have much work to get done.

The G5's can be called desktop machines because they have a desktop OS installed. Solaris does not count. Windows would count if their 64 bit OS wasn't specifically intended for servers. Linux would count if it was an OS for the end user (it isn't- yet). Desktop machine doesn't mean that it fits on a desk, it means that it can be used by end users and is readily available to end users.

You can't allocate more than 4GB of RAM to any windows process in 32 bit mode, so having more than 4GB of RAM is worthless to any program.

You are overlooking the real problem, and that is that the only upgrade path for 64bit in windows land is for servers. This is a MAJOR problem. I write windows software for a living. If Intel came out with dual 10GHz chips that could address 16GB of RAM, we would reccommend it to our users. The problem is that all 64 bit computing in windows land is stuck in server mode and this really stifles the migration to more advanced processors for your average software developer. Do you think that in three years that we will all still be content with 4GB of RAM? Apple has taken the plunge to 64 bit computing while MS is going to repeat the heinous migration that came with the transition from Win16 to Win32.

Video editing can sometimes benefit. Consumers aren't using >4GB in photoshop and certainly not ripping CDs.

So we agree that a lot of workstations need more memory, consumers don't really(yet).


If Panther is 64bit, then it will probably beat the systems with consumer winXP + low-end AMD Hammers to market. Last I read, they were 4-5 months off.

Except that you can get Itanium2 workstations with 64bit windows(for workstations) at ~$3000. Itaniums have been out for like 2 years and the opteron has been out maybe a month. It seems MS is supporting the new chips adequately, I don't see this stifling you keep mentioning.

Now I don't write windows software for a living(CS grad with no experience. ugh), and I've never written anything that uses a lot of RAM but I believe there clunky ways to address >4GB. At the least you could coordinate multiple processes.

I think that a lot of the problem with the benchmarks from both Windows and Mac Users, is the issue of Apple in the past pushing Altivec performance as being representative of the G4's performance, which it was not overall. Now I think a lot of people, just can't except the fact that a new Mac exists that doesn't have to have its performance "inflated", so to speak, to be marketable.

Reality, is that there are bound to be some areas where Wintel or WinAMD machines will perform better. (I bet they sroll through word much faster) However, for Apple's key markets, and in overall performance across a wide spectrum of applications, I think the G5's will more then prove their worth, and will excell within their key markets.

I kind of suspect that Steve isn't really too worried regarding all the excitement about the benchmarks. Most likely he expected it and maybe even encouraged it a little. In case no-one has noticed, he has mellowed a little with age. He can laugh and make public jokes about "mistakes", and while always the perfectionist, can grudingly accept other folks limitations to a small degree. I just get the feeling that he is kicked back chuckling, knowing that the full power of the G5 hasn't yet been revealed. Knowing that what he showed us was only the beginning. Enough to get the attention of the markets that he is concerned with, and enough to stir folks interest and ire. After all, a little controversy makes for some awefully good cheap advertising.

Once the first G5's are available, and in the hands of reviewers, you already know that they are going to be put through every conceivable test possible and compared in every way possible to the best that Intel, AMD and Microsoft have to offer. Before summer ends, the G5 will be almost a household name. Then will come Panther, and it will start all over again. And as each test, each comparision, each evalution both pro and con comes out, Steve will chuckle a little louder, knowing full well that when the dust finally settles, there will be a shiny aluminum box with handles and an Apple logo, sitting at the top of the pile.

Can you imagine a day, not too far off, when PC reviewers instead of running comparisions of Intel to AMD and AMD to Intel, will instead compare the latest and greatest from those companies, to the G5? Can you imagine...???

I can IMAGINE it!

And secondly, Apple's G5 dual 2 gigger is now 3rd on Apple Store charts...it's stormed the charts, gunning for the iPod in sales!!!

Amazing. Pent up demand from all those G3/G4 users?

I think Apple's quarter's are in for a shot in the arm!

I think the G5's performance was undersold by focusing on Spec marks. I think this was a masterstroke in sharp contrast to the over egged G4 benches which depended on Altivec to look good. This time...Altivec is merely the cream on the cake. It's not the crux of the G5 system. The G5 is awesome in everything! FSB. Memory. CPU. FPU. Integer okay. Huge bandwidth... Altivec on bigger bandwidth. Wait until reviewers DO test this beast. Mac press will go: 'Wow. If you were G3 owner? Upgrade. WOW. If you were G4 owner? UPGRADE!!!' PC press will be stuperfied that a Mac can cream PCs in Intel biased Spec. Then be stuperfied against when G5 Photoshop and Lightwave benches cream their best workstations for over a grand less. And we're about 6 months from the next bumps and PRICE CUTS based on last year's rather 'slow' G4 updates.

Brad's Motioncapture software really showed a G5 PPC stomping on the PC workstation's home territory! It was far more fluid. Eat that, XEON! Eat that, DELL! More over, Wintel, there's a new Workstation Desktop PC in town!

I'd like to see Apple intro' new speed grades as they become available and bump the line regularly rather than waiting 6 months. 2.2 or 2.4 or 2.5 970 by Fall?

If the G5's are cheaper chips...then Apple may make a killing on the dual 2 gigger...maybe madness in the single G5 Tower pricing...

3 gig inside a year!? That's got to be a bump at Jan' at the latest? If the G5s at WWDC ship later...it may be worth the wait for a bumped machine with Panther installed...

I don't know if I'll wait that long! Drool!

The G5 is kickass!

Lemon Bon Bon