MacBidouille posts PPC 970 benchmarks - Page 6

His argument is simply that at the time of the 970's initial introduction it will probably be at around 80% of the top-of-the-line Intel performance of the same day, based on the facts that we know about the 970. There are no facts saying that there will be a 970 at >1.8 GHz so arguing that a 2.5 GHz 970 will do better is currently just wishful thinking.

yes. i'd bet 2.5 is a Jan. thing. Yet another reason for apple to adopt an AMD type rating system.

Powermac G5 3100

only 45 ish days till we all know.

Well there was that leaked file from IBM stating up to 2.3GHz 970's going to be ready. No one can say what speeds Apple will get from IBM, 1.2, 1.4, 1.6, 1.8, 2.0, 2.3, etc. It could be anything, its all just speculation at this point.

But the thing that everyone seems to miss (or maybe not, and I am just way off!), but look at the current P4 compared to G4. Loking at SPEC the P4 at 3.0GHz scores 3-4 times as much as a G4 at 1.25GHz (both single). However, in some tasks the G4 can be somewhat close to the P4, and in altivec enabled tasks the G4 can beat the P4.

So now looking back at the 970. We have a chip that has estimated SPEC scores pretty damn close to the P4. So I would bet that the performance difference in benchmarks will put the 970 much close than 80% of a P4.

And I mention estimated SPEC, because if you look at the Opteron as an exmaple, the actual SPEC scores were better than the predicted. This could happen with the 970 as well.

So what am I saying? Before everyone decides the 970 sucks, is underpowered, would get beat by a P4, (insert your dissatisfaction here), lets wait for it to show up. On the other hand, don't think that the 970 will destory a P4. Lets wait for it to show up

Is that two DDR signals 180 degrees out of phase, like the PIV's bus?

Why do we keep insisting in the cross-platform benchmarks? You know, the ones that have been flamebait since the original PC's intro back in the dark ages.

One thing no one disputes is:
A dual 970 will obliterate a dual G4.

So no matter what the (estimated/rumored/etc) numbers say precisely... we're going to be a _lot_ better off.

I think the reason for it is that in some fields ( 3D ) you really have to have the fastest setup available at all times, and there is no skimping on your hardware. I'm waiting to until these are released with benches before I make my next purchase, but I absolutely have to know how these machines stand up against a Dual, and Quad P4 Xeons. I need render stats, and that is a huge part of what I'm basing my next purchase decision on.
It's either going going to be a dual Xeon, or a dual/or quad 970.
I've never owned a PC, but now I have to choose based on speed, and I have to accept it.

Gamers have a similar argument. Game developers are also wondering if this could be the ultimate game machine to produce games for, and if the customer base will be there. You know they are already thinking about how, and when to tackle 64-bit issues with the next Gen. AMD processors. Apple, and AMD could be a big reason to switch most efforts over to a 64-bit aware architecture, and begin the next generation in gaming superiority. (gaming performance, and availability on the Mac is a huge issue with PC user's to)

So many PC users hang on to the PC, and stay away from the Mac for speed reasons alone. Face it the Mac has become second rate hardware in comparison to what wintel has at their disposal.

I know a lot more than 20 PC users that love OS X, and would jump the POS win-doze ship in a heartbeat if the hardware was in the same league, but it isn't. And they swear off buying an overpriced Mac at lesser performance when they can get a much less expensive machine that outpaces the fastest Mac in all areas.

I think the PC "cross-platform" performance comparison is a VERY relevant issue.

I don't think the "hardware is in the same league" argument cuts it. The PowerPC platform is in the same league. It might be slower than the fastest P4, but for normal use, people won't notice the difference. PC people might give you an excuse why they won't jump, but for most of them, it's not a rational reason.

PC gamers will never switch over to the Mac, because of availability of games and benchmarks. But who wants PC gamers, they are a small and useless market.

Macs aren't for everybody. The introduction of a new CPU won't change that. But it will still be a great treat for "the rest of us."

Onlooker, I agree.

For those PC gamers eyeing 'X' covetously I think the 970, especially in its dual form, will provide a compelling switch argument. Never in recent history has the Mac become such a game inundated machine. You can get most of the best games. A dual 970 with Radeon 9700 is going to be kick-ass for games. The 970 removes alot of barriers to performance eg bandwidth and that extra fpu...and performing twice per clock as the G4 or Pentium 4 does will send shivers down Wintel spines.

And Photoshop and Lightwave performance (Hello LBB!) will be quite something else, I'll bet.

eg, say...1.8 gig to 2.5 gig 970s come Jan 04? That's easily a 5 gig G4.../Pentium 4. I don't see Intel scaling that fast in performance.

I suppose it's easy to get carried away with 'paper talk', however, a point above was well made. The 970 will hammer the G4 and leave current Mac-heads who are gasping for decent 3D performance with a reason to choose PPC 'X' rather than Xeon 'XP'.

Note the absence of Moki. Quietly confident he. Apple's management seem so too.

The glass and chrome desk is ready. My credit card awaits.

Lemon Bon Bon

Be ready to go broke

I have to agree and to disagree.
Agree: Games will mostly be written for the 95% market share of windows and perhaps then be ported to the mac. They will be faster on a wintel system because they will be optimized for it. Only games with a big mac comunity will be optimized for mac os (and then be faster than the pc version )
Disagree: PC gamers are a big market which drives the the whole GPU development.

BTW: SPEC became more and more intel friendly from version to version over the past years. So if the 790 gets 80% of the rate of the fastest P4 that doesn't bother me - I'll buy that machine!

[QUOTE]Originally posted by onlooker
I think the reason for it is that in some fields ( 3D ) you really have to have the fastest setup available at all times
A DP 970 will be one of the fastest desktop machines on the planet.
but I absolutely have to know how these machines stand up against a Dual, and Quad P4 Xeons.
Quad P4 Xeons? You must have money piled up to your chin to buy a system with 4 Xeon MPs

And about that 64-bit awareness, like it was mentioned before (Programmer ?) it's more about the power these new systems have because they are a new generation of chips with more muscles everywhere than just having 64 bit integer units.

spec is actually quite a bad benchmark anyway.
at least I always think so.
It's surreal, compares poorly over different architectures etc...

good idea, badly done.

To accomodate PC gamers, Apple would have to destroy the integrity of its tight integration.

They won't even buy Dell or HP because that doesn't offer them enough flexibity to upgrade CPU/mainboard/graphics card every 3 months...and you think they will switch to Apple? I don't think so.

Chasing this market is pointless for Apple. Sure, it's nice that the Mac has games, but that's not really its purpose. The console gaming experience is much more simple and hassle-free than using a PC to play games, and the Mac is much more simple and hassle-free than using a PC to get work done.

There are some interesting AMD benches out there now. Looks like AMD's 64 and dual 64 bit offering SMOKE the latest Xeon and dual Xeon setups to the tune of beinf twice as fast.

Looks lik PPC970 has some competition comming it's way. 970 does have the advantage of being just cool enough to go into a PB though!

That post went unanswered for days, but I stand by what I said. All I can say now we'll see what happens when our hardware becomes comparable to that of wintel.


I totally disagree. It is not in the same league.

Motherboard on a G4 is not equal to that of available PC motherboards.
here is a list of what they can get on one motherboard that we can not that ROCKS!

• DDR 3400
• Serial ATA
• USB 2.0
• PCI X
• Superior hardware rendering Graphics cards for 3D
• MAYA 5.0 UNLIMITED (ok it's not hardware, but.....)

Don't forget the P4, and the P4 Xeon are wicked fast processors. Just one more thing to pile on that list.

That list takes all the Power right out of the word PowerMac.

Heck, the DDR that comes on current PowerMac's doesn't even work properly. From what I read the G4 isn't designed to use it right.

If anyone thinks "the hardware isn't in the same league argument doesn't cut it" they are only fooling themselves.

The PowerMac is lagging sooooo far behind it's pathetic.

Boy did writing that suck. \

That is exactly the sort of short-sightedness that will kill the Mac as a viable platform.

The G4's have been behind the curve for many common tasks, though they still excell at a certain few other tasks. The high end P4 and AMD's all excell in most general tasks (though each has its strengths and weaknesses) but fall behind the G4 on certain others. Overall though the G4 is behind on most common tasks and no one with any sense is going to argue that point, and the margin by which it might excell on any specific tasks grows smaller daily.

The G4 however is for all intents and purposes, history. The pertinent argument is the potential impact of the 970. The 970 changes everything. It changes the type of motherboard architecture Apple can offer, it changes the types of enhanced technologies that are available. It changes even the way that Apple will think, work and market. From the Macintosh perspective, it more then just evolutionary, it is revolutionary. In one gigantic step, a whole new world of possibilities opens up.

Will it blow away the Pentium? That is a matter of opinion. I suspect it will greatly outperform the Pentium on a great many types of tasks, and at the same time there will be some tasks where the P4 will still excell. Overall, I suspect the edge will go to the 970 for a little while. Likewise the AMD 64 bit chips will have their "areas" of excellence. Until we get real world working machines though, none of us can know for sure what any of them can do. The Power PC 970 Macintoshs will probably be designed to excell in the graphic and creative markets, along with certain scientific markets. They well also be overall well rounded machines, that are easy to use, fast and reliable. They dont' have to be the fastest at everything, they only need to lead in their primary markets.

As to gamers? Well this is a huge market. The very hard core avid gamers who push the limits and drive the technologies, compose probably no more then the top 10 or 15% of the gaming market. These folks like to do it all themselves, and build their own machines. The Macs will probably never appeal to them. However as you decend the ranks, at least 50% of the gaming market is composed of casual and semi-avid gamers. They enjoy the fun and competition, but really don't want to deal messing with their computers beyond simple upgrades like RAM or Graphics cards. For these gamers, any Apple computer that is competitive may be of interest. The gaming market would probably be more of a secondary market for Apple, but make no mistakes it will not be an ignored market. Much of the technology that appeals to the graphics, creative and scientific markets will also be applicable to the gaming market. Not the top 15% of the gaming market perhaps, but very easily at least the lower 50-60%, and that still translates to hundreds of millions of dollars. With even as little as 10% of the total gaming market, Apple makes money and gets an excellent return on investment. With the 970 they can probably do better then 10%.

What has kept Apple afloat through the last few years, hasn't been their hardware, it has been their OS. Hopefully soon that OS will soon have hardware truly deserving of it.

Again I say, forget the G4. It is ancient history. Its future lies in the embedded markets, not in computers by Apple or anyone else. The Xeon's and Athon 64's will no doubt give the 970 a run for its money on raw performance, as they should. However in the end, each excells in its own arena, and there will be no one champ, though all will provide a wide spectrum of good performance. Competition will drive continued improvements The final differentiating factor will be the OS, and that is where the Mac has always shined, and will continue to shine above the others.

There are a few things that are really advantageous for Apple right now.

1) Because of the economy, there are a lot of people who are sitting tight on Wintel equipment that DOESN'T blow the current Macs out of the water. There's been an increasing gap in speed the last year or so, but at the same time, people are finding themselves satisfied with their current rigs and are not upgrading like they once were. Strangely, for that reason there are probably a lot of people who are unaware or unphased by the Wintel-Mac gap. They're not thinking about buying a new computer yet...they're still trying to figure out how to make their Dells and Gateways perform without constant viruses, worms, and the pitfalls of OS Decay. It could very well happen that by the time they are willing to buy, most consumers will have skipped right over this graceless period in Mac history without understanding how far behind we were.

2) How many versions of Windows are there in current popular use? Five? Every one was supposed to squash the bugs of the one before it but never did. MS says that XP is 'it', but they said that about Windows ME too. My Windows bretheren are losing faith and feeling mightily rooked by MS. The draconian licensing scams are an added insult. Most folks I know who actually do the buying just want a 'good' computer that will allow their kids to do their homework, connect to their digital cameras/camcorders, burn and listen to music, and allow them to browse and do email without screeching to a halt for no apparent reason or getting infected with viruses or spyware. Windows is poor at multi-tasking without crashing, and rare is the crash that doesn't take down the whole system. As my good friend said about my last Wintel computer (and he uses Wintel), 'Yup, crashes twice as fast as your old computer. Wow! Watch how fast those error messages come up!'

3) Apple's been fighting with its hands tied behind its back. This means it has adapted in a desperate attempt to keep up with inferior chips. This has brought about, for instance, adjustments to the OS and some of the leading software that makes them much more multi-processor aware than Windows. Every scrap of design advantage necessary to try to compete with a lagging processor and a strangled FSB and memory utilization means that that efficiency is ready and waiting to serve any new processor and motherboard just as well. There's no sloppiness of expecting the processor speed to make up for otherwise poor design.

Granted the Spec scores, duals, and the bare-minimums on the ppc970, where would we be? 'Serial ATA' and 'USB2.0' aren't issues anyway. SerialATA's cool, but the market for Dual Xeons seems to focus on FiberChannel or SCSI.

We have no idea what memory the 970 will actually use, but it seems like the duals will need two banks to stay saturated -> a plus for the 970 if Apple springs for the extra mb cost. Apple hasn't historically used memory that won't saturate the FSB on their top end machines.

The current ATA133 compares pretty well with some SerialATA implementations (which is 'ATA150'). Single drives don't dump >100MB/s. RAIDs can use Fiberchannel.... So, this is a bit murky. I don't see why the 970 wouldn't have SerialATA onboard , and I wouldn't panic if it didn't.

FW800 onboard, USB2.0 (at the _very_ least) as a 3party option. I'd say that's a plus for ppc-970. Or at least a neutral.

PCI-X. I'm not sure where they're going here. The current XServe has two 64-bit 66MHz PCI slots (plus a third combination PCI/AGP slot). Not shabby, not 100MHz PCI-X though. I don't see why there wouldn't be at least one PCIX slot though - Apple's PCI implementation has always been updated at major motherboard revisions.

GPU. There's a couple of people nattering about GPU/hardware rendering/whatnot. _I'm_ content it is being addressed. And the way the graphics folk keep chuckling, I'm wondering what else might be going on here.

Maya... well, that's not owned by Apple. But Lightwave should get a _major_ performance boost over the G4 (due to extra FPU). And Shake will be there... Reason... Plus whatever Steve's couple of 64-bit 'trick pony' apps are going to be (Oracle?). Since there almost has to be _something_ to show that the 970 goes to 11.

So as far as "motherboard features" are concerned, I'm not worried. When the _current_ MDD PowerMac's motherboard was designed, not many of the listed things were all that common (other than USB2.0 let's say).

It's been a year, they aren't going to use the same freaking thing. It _has_ to change completely _anyway_ for the new CPU/companion chip/RAM combination.

You know, all this "tight integration" talk sounds very nice, but the reality is often quite different. Apple has at least its fair share of problems with supporting the hardware it has specified and you have very restricted choices of what hardware to use. The reason that some of the people stay away from Dell and others is that they have implement proprietary standards which restrict their freedom of choice of hardware.

The argument that gamers are a small and insignificant market sounds like the argument used by most people, except the word "Apple" is inserted where "gamer" is in this one. The simple truth is that the gamers are at the leading edge of hardware development. A recent survey showed that the gamers have what is, on balance, the best hardware of any identifiable segment of the PC market (specifically referring to graphics processors and CPUs, but it also applied to most of the rest of the hardware). (The industry loves them if no one else does.) It is not that difficult to let the manufacturer write the necessary device drivers for their products if you make your system capable of using standard hardware. The manufacturer runs the device driver through the OS compatibility certification process and gets to use the good housekeeping seal of approval on their product. What's so bad about that?

The problem is that reality frequently does not measure up to Apple's claims about hardware. The frequent discussions about an X86 version of OS X reflect the frustration that many people experience in dealing with the restrictions of Apple hardware and Microsoft's OS (and licensing practices). There is increasing dissatisfaction with Microsoft's operating systems in the PC community. If Apple were to provide an alternative OS for the existing hardware who knows where it might end? In the end computers are simply tools, whether they be PCs or Macs. The problem that Apple has is that its tools are typically more expensive than a comparable tool from the other side and so there really is no comparison. Unless Apple can demonstrate a cost efficient benefit to its tool it will not succeed in the larger market. Recent sales data show that Apple sales continue to decline in market share. At some point the amortized costs per unit sold will climb and the rising price will adversely affect sales which will.... Apple must get the prices of its hardware down to be competitive. Perhaps not down to the exact same price, but something in the ball park that will at lease cause people to consider it as a possible choice.

Any truth to my instincts that the programmable GPU (obviously copuled with a faster CPU/Motherboard in general) is going to help Apple's image as a gaming platform?

If this "small and useless" market that tend to buy expensive computers and upgrade them a lot comprise of say 1% of the PC market that is equal in volume to 30% of the Macintosh market. If even a small fraction of those would go to Macintosh it would be a very very big thing.

Listen, I know a ton of "l33t g4m3rz." They are NOT a part of the PC market. THEY DON'T BUY PCs.

They might buy motherboards, RAM, CPUs, sound cards, graphics cards, and power supplies, but they DO NOT BUY PCs. They have a huge investment in their hardware and software, and they are least likely to switch to Mac out of any major PC user market.

How could Apple accomodate them?

1)100% fullspeed compatibility with all existing Windows games. They wouldn't switch if they lost any of their old games, or even if one new game was Windows-only.

2)Full upgradability for PowerPC machines, including drivers for esoteric hardware, motherboards and CPUs available a la carte.

If both of either one of these came to pass, they would completely obliterate the Mac platform, leaving no one with a compelling reason to buy an Apple (besides aesthetics).

PC gamers are a fickle, whiny, and contentious lot. They spend a lot, sure, but they don't spend a lot on anything that Apple provides.

You may as well argue that Apple should start targeting baseball card collectors, because they spend a huge amount on their chosen hobby. It makes as much sense.

From IGM

I don't think gaming is as inconsequetial as some pretend. The hardcore, which I don't understand, will buy the stuff first, when it costs the most, but a lot of other people buy the stuff later on, as it drops in price, and they use it to play games. The market is big enough that everyone in hollywood notices, your politicians notice, and market forcasters/analysts brokers notice when, for example, nVidia is set to release nv-35. That equates to making them a play of the day on the basis of gamer response to the product, and the inevitable trickle down. "Gaming" is more than just the hardcore, gotta have it first set, it's kind of a default performance branding that moves a lot of parts into consumers homes in the months that follow (whether they "game" or not) It actually has a very profound influence on the market. The only single aspect with a deeper influence on tech and communication, is porn. The unofficial "app" that trumps both, is of course, "sharing."

Hmm, I wouldn't bet on it

All that means is that they'll honestly attribute the report to rumors if that's their source. And no, that doesn't mean rumors from professional contacts. That would be attributed to "undisclosed sources" not to rumor.

What I want to know is: which one of you MacRumors readers works for Forbes? C'mon, fess up!

a rumor reported by a more respected online magazine is still a rumor.

Is it June 23rd yet?

Do you have any quantifiable evidence here? This sounds like the same kind of bull-AHEM-hot air that we used to hear about the "information superhighway" ten years ago.

Your "trickle down" argument is weak. You could say the same thing about any intensive computer task, 3D rendering, video editing, etc...

Gaming doesn't drive any market besides the gaming market. Which, in case you haven't noticed, is in a deep recession right now (especially PC gaming). And did you just say that porn has a "deep influence on tech and communication"? ("I'm just doing...research, I swear!" ) I *might* accept the converse, but outside of developing increasingly annoying web ads, porn has no influence on communication or tech companies.

Let's not dress things up here. Benchmarks (and gamers) have much less influence than most people on this board seem to think. Price differential and misconceptions about compatibility are the the two main things holding back the Mac.

The PPC970 is a great leap up for traditional Mac users and techie switchers who are looking for a powerful UNIX workstation, but it's not gonna cause any mass migrations from Windows and/or the x86 world...

In 1997, when Intel was shipping Ppro-200s to Apple's dual 604e-350s, you can bet there weren't any PC users burning up message boards saying that x86 was dead, and that they could never catch up. Relax, the CPU issue is not as big as everyone seems to think it is. And whether these benchmarks are true or not, it will be resolved soon

Gizzmonic: well said.

Gamers are important, if only for PR and bragging rights.

In case anyone is interested, here is another pdf file from IBM on the ppc970.

link

The last sentence in that MPF Review you cited is my favorite.

Interesting read. It seems those who are dreaming about a 2.3ghz 970 are well, dreaming. The pdf states that 970 max speed is 1.8ghz. So, that will be the top end, 1.8ghz.

I for one will still be giddy with a 1.8ghz this year.

Note that it was written in October 2002.

AFAIK Apple never released a dual 604 350MHz. The fastest MP machine they released back then was 200MHz.
My 9600 350MHz was made in November 1997, during the same year Intel released the Pentium II up to 300MHz. So the speed differences were not as great as you make them out to be, not at all comparable to the differences between Motorola and Intel CPUs today.

Not saying your wrong or anything, but the date on that pdf is 10/28/02. That was written shortly after the Microprocessor Forum.

A lot can happen in 8 months. Like maybe the inadvertent press release by IBM Germany of up and coming 2.5ghz blades.

Oh, by the way, I also "will still be giddy with a 1.8ghz. this year".

opps, Programmer beat me by 3 minutes, I must learn to type faster

AH, yes, entertainment has done more to boost technologies than most any other form of promotion. VHS-Video? Porn. 1-900? Porn. The internet? HELLO! It took quite some time from the internet going "commercial" before any of the top ten money earning sites was anything other than porn. It is a killer app, and there's no disputing it.

Gaming may not be a killer, but it counts, the market for games is huge, the consquences of presence (or lack of it) are also profound.

If Gaming weren't high on the radar, it wouldn't be all over CNN or NBC's news channels whenever ATI and nVidia release a new product. We're talking main stream media here (as much as financial news can be considered main stream) Would anyone pay attention to Sony's PS2 fortunes or M$'s Xbox forays if gaming wasn't important?

And where exactly did all of nVidia's and ATI's embedded/OEM products originate? They trickled down from cards marketed directly at GAMING! The virtual dissapearance of Matrox can be traced in lock step with their failure to capture attention in the gaming community.

I'm not saying that Apple has to cater to gamers with a completely modular architecture, I'm saying you simply can't write off an area of the industry because "that's just games." There's no such thing. "Games" make more money than hollywood box-office in many parts, and they do drive technology. Does your average office PC have the latest GPU? Nope. Gaming rigs do.

In the end, the argument is not weak at all. You cannot say the same thing about any intensive computer task, 3D, A/V, modelling, yadda yadda. How many household computer users do those tasks? Not many. How many "share" files, play games, and surf pr0n? Just about every other one. Which one drive the market again? And since most consumers are more likely to play doom than they are to model anything, it is games that establish a consumer brand and give it market potential (as far as GPU's go). Those tend to end up business models, in schools, as well as, in homes.

Nowhere, you'll note, did I say that PPC970 would bring droves of gamers over to the mac. They'd be foolish to do that, Apple would be foolish to think it.

I'm just saying that you can't underestimate games. It tends to be a familiar refrain around here, this idea that "games" should be dismissed because they're inconsequetial. They aren't, not in the least.

And I will be Ultra Hyper Giddy with my dual 2.5 970

Also says a lot can change before the chip comes out and chip-sampling 2Q 2003.

Ok, so we know that the 970 is eventually going to get faster than 1.8 ghz, but the 2.5 blades were supposedly prototype 0.09 parts that supposedly will be introduced later than the inital 0.13 ones (early next year?). So, was this something they said to save their german butts or is it possible that the 0.13 eventually scaled a lot better than IBM had initially hoped for. As you said, a lot can happen in 8 months. Heck, even the Power4 has reach 1.8 now, and as far as I know that is still at 0.18 (right) with a shorter pipe-line than the 970. Yeah, I understand that the Power4 has been in production for years now, and they probably have a good understanding of how to improve its speeds, but can it be possible that this knowledge in someway can be inherited by the 970. Just thinking out loud here