IBM PowerPC 970 Architecture

riclewis writes "Hannibal from Ars Technica offers an explanation of some of the internals of the new IBM chip. It's certainly more powerful than anything on the desktop now, but by the time it's released a year from now, it looks to be middle-of-the-pack (which could still be a step up for Apple...) This excitement over the early release of hardware specs kinda reminds me of all the hype surrounding the Sony's Emotion Engine when it was introduced a couple years ago. In fact, some are suggesting the PPC 970 chip might be closely related to the PS3's 'Cell' processor..."

Comments

PS3 apple?

[barfarf]

"In fact, some are suggesting the PPC 970 chip might be closely related to the PS3's 'Cell' processor..."

Even though it's really doubtful, it'd be extremely cool to see a PS3 emulator on the mac if the processors are that closely related.

I remember running Mac OS 6.0.5 on my Atari ST. Because it had the same processor, it didn't need much to make it run.

Oh well, I can at least dream, can't I?

SPEC INT and SPEC FP numbers (937 and 1051)

[WittyName]

Not that hot NOW! They will have a lot of competition in that space with Opteron/Clawhammer, and the new Sparcs.

Still, glad to see something other than incremental progress.

Re:SPEC INT and SPEC FP numbers (937 and 1051)

[Anonymous Coward]

The problem with this is that is both compiler dependent *and* OS dependent .

Well, that of course depends on what you're interested in.

Since we use computers to run and compile our own code, this is EXACTLY what we want. The vendor can do anything but touch the code. If they release a compiler that produces twice as fast code they'll get better benchmarks, and that will show up in our performance too.

As has stated many times before, the current G4 machines score in the low 300s in SPEC marks. Does this mean that the G4 is 3 to 5 times slower than the P4? In practice, it isn't.

In some cases it defintely is. Have you compared compile times with gcc on a G4 and an x86? It is horribly slow on the Apple box, and this is reflected perfectly in the gcc benchmark of SPECint.

Of course SPEC benchmarks aren't 100% accurate, but a lot of people seem to believe that they are unfair against Motorola. They aren't - but they don't make any claim whatsoever to measure performance of code that has been handtuned with Altivec or SSE. This means you can get excellent photoshop performance on a G4, but it still sucks on general-purpose compiled code as long as there isn't any compiler that can generate Altivec automatically.

Now, if you only run a small number of Altivec-accelerated applications (as many Mac users do), that perfectly OK. But for scientific stuff that we do SPEC is a very good and impartial indicator of performance.

Oh look

[Mononoke]

Lookee here: It's another story containing flamebait:

It's certainly more powerful than anything on the desktop now, but by the time it's released a year from now, it looks to be middle-of-the-pack (which could still be a step up for Apple...)

Ya know, if you can't tell the difference between chip architectures by now, there's no point in trying to explain.

It's a shame for apple that IBM announced this.

[bluemilker]

I mean... it's great news that Apple won't have to rely on Motorolla's decidedly passive desktop chip development strategy anymore...

But man. First off, this kills any possibility of a big surprise hit. Second, this dooms apple sales for the next year or so... who wants to buy a stagnating desktop model when the next edition has so much promise?

Then again, Apple's desktop offerings have been a little stagnant anyway... most people probably won't want to play the waiting game for as long as it'll take for these to come out.

I just hope that by the time they do, they're worth it.

Re:Apples and oranges

[Christopher Thomas]

As has been well-documented, Macs perform just as well as Windows machines. The slower clock speed of PowerPC compared to Intel is made up by the lack of code bloat is Mac OS compated to Windows.

A couple of points to throw water on this:

• Not all x86 machines run Windows.
  
• Windows "code bloat" mainly slows down the interface, as opposed to number crunching, which is where you'll be processor limited. It also eats memory for breakfast, but if you're doing anything computation-intensive you'd *better* be throwing memory at the machine no matter what.
  
• Apple became allergic to producing useful benchmark comparisons some time ago. Find me SPECmarks on an actual Apple machine submitted within the past 3-4 years. All you find are photoshop benchmarks.
  

Apples are certainly wonderful machines, and Windows certainly is icky most of the time, but be prepared to back up any benchmark statements with actual benchmarks.

Also, PowerPC and Intel/AMD are two different types of processing, so they can't really even be compared.

Um, no.

All general-purpose microprocessors perform certain basic tasks upon which everything else is built - integer and FP math, memory access, and control flow operations. Processors take different approaches in how they implement these functions, but the interfaces presented to programmers - even assembly programmers - are very similar [and yes, I've done assembly on multiple platforms].

You can also completely ignore architecture and take test programs that you think are representative of the kinds of tasks found in different types of application, compile them for both platforms, and measure how long it takes to do the same amount of work on each machine. This is the _foundation_ of benchmarking.

If the machines were completely different, you wouldn't be able to do the same tasks on them!

Re:Apple Chips

[CTho9305]

Everything you said is correct... BUT...
It doesn't matter how much work a processor does per clock, if you can scale an "inferior" (according to your definition of inferior) to a MUCH higher clock.

This may not even be an architectural flaw as much as the result of an inferior manufacturing process. If Motorola's fabs aren't as good as Intel's (I don't think they are) then the fact that the G4 is a "better" processor on paper is completely irrelevant - for all the consumer cares, the FASTEST G4 available is slower than the fastest P4 (Currently, according to benchmarks not done by apple, it seems that you dont even need the absolute fastest P4s to beat the fastest Macs)

Re:Specmarks similar to Pentium 4 2.8 GHz

[Hadlock]

yeah, but a) it's running 1GHz slower than the P4, and b) i doubt it's dealing with ultra-long ints (or is that the point of FP? i dunno..)

Re:Apple Chips

[Don Negro]

Does it?

The eetimes story linked at the top says it's an 8-stage pipe. That doesn't mean any more or less than the extreme tech statement that the new pipe is triple the length (which would be 21, the current pipe is 7) since we haven't seen any actual reference docs from IBM.

Can anybody who was at the Microprocessor Forum give us more info?

Re:Is this a joke?

[Anonymous Coward]

I see that you don't actually use macs, you just mock them. If you used them, you'd realize that the real world performance of a new bottom end mac desktop (a dual 867mhz g4) is both price-competitive, and performance competitive with a similar PC (such as a middle-of-the-road Dell Optiplex)

not to mention the fact that OS X provides ease of use that windows, linux and bsd can't touch, plus it has all the software you need to get work done, whether you want free software, or commercial software.

Re:This is gobbledegook to me...

[Anonymous Coward]

My reading is that there are four significant differences between IBM's chip and the G4.

• Clock speed. This one is straightforward. While comparing clock speeds between different types of processors causes confusion (a Pentium clock isn't the same as an AMD clock isn't the same as a PowerPC clock), higher clock speeds in the same processor family are always better. When it arrives, the PowerPC 970 will max out at 1.8 GHz compared to the top speed of 1.25 GHz of the G4 today.
• Instructions per cycle. Here's where things get tricky. G4s execute a maximum of 3 instructions per clock cycle. The PowerPC 970 will be able to execute a maximum of 8 instructions per cycle. Before you get too excited, consider that those are maximums and in practice it may not be possible to execute that many instructions. Due to changes in the chip architecture, it is more likely that not all possible instructions will be executed. Still, this should result in a speed improvement.
• 64 bit addressing. G4s are limited to 4 GB of RAM due to their 32 bit architecture. The PowerPC 970 has a 64 bit architecture, so they can support up to 4 TB of RAM. (I think. I'm a bit hazy on what actually controls the upper limit.) This means that graphics applications that require absurd amounts of memory will have more room to grow. The 64 bit architecture also means more computational precision, but it's unclear to me how useful that would be outside of scientific computations.
• Bus speed. The PowerPC 970 supports a 900MHz bus, which is much faster than the G4. This controls the rate at which the processor can access memory. For memory intensive applications, bus speed can be more important than processor speed, because the processor ends up having to wait for data from memory much of the time.

My expectation is that the bus will make the biggest difference for end users, followed by the improvement in instructions per cycle, at least in the short term. Then again, I'm far from an expert, so someone else might have better understanding of the potential performance gains.

Matthew

Friends don't let friends Slashdot.

Re:Apples and oranges... and dollars

[Smidge204]

It's often asked -- How do you compare apples and oranges?

By nutritional value, of course!

I't hard to really come up with a good, fair way to compare two different chip architectures. Even using the same program written for both doesn't indicate a true comparison between them overall.

So it seems to me that the best way to compare is popularity. I realize that's not a terribly fair method either, but since Apple seems to be putting a lot into the next generation of machines, success comes from improving their market share (I know, 'duh'), and that will only happen if the new machines are really up-to-snuff in the public eye.

So you can argue about how macs are better/worse than IBM machines 'till the cows come home. What it really comes down to is if the new machines will let Apple can break out of it's pseudo-nitche market.

=Smidge=

Osbourne effect doesn't make sense

[Sloppy]

with the 970 looming on the horizon and the G4 apparently stuck again around the 1GHz mark, nobody in their right mind would shell out for a new PowerMac any time after mid-2003.

There is never a good time to buy a computer, and nobody in their right mind will ever buy one at all. There is always something faster coming up.

Once you get over how ludicrous that is, I say buy a computer whenever the hell you want one. And yes, your machine will be obsolete, according to all the charts and graphs and tables of benchmark numbers, almost immediately. It doesn't matter if you buy a G4 in 2003, or a 970 in 2004. It will still happen. Get over it.

Re:Chunks of five

[Courageous]

computing in chunks... sounds a lot like a Cray

This chunking is described in great detail in the original POWER4 public design documents. It's referred to in passing as a redeeming feature, borrowed from VLIW concepts. The suggestion is that its a part of traditional VLIW that could be leveraged into a non-VLIW design.

C//

Re:Chunks of five

[Scott Wood]

Everybody, that is, except the people that want *good* optimizations, rather than merely "sick" ones. Take a look at the code GCC produces sometime... it's not pretty.

MP?

[muchmusic]

I may have missed a comment here, but I'm not that familiar with this subject. Apple uses multiprocessor schemes in all of its pro desktops now - I know that it's mainly to make up for other speed deficiencies - but is it possible (probable) that we will see dual processor versions of desktops with this chip as well?

what other option is there?

[GunFodder]

How do you make a CPU go faster without jacking up the clock? You can create a new design, but that would be a lot more expensive than jacking up the clock. It takes years to develop a new CPU design and IPC increases are usually minimal. Most new CPUs seem to be designed to run at higher frequencies to achieve better performance. The P4 actually has a lower IPC than the P3 but can operate at much higher frequencies.

Many companies are planning to use parallelism to improve performance. IBM has a CPU with two logical cores, and Intel will introduce CPUs with two virtual processors in the very near future. But parallelism is not likely to get you a doubling of performance, especially on a desktop machine that is often running only one intensive process at a time.

Apple is OK, Mac user's aren't performance freaks

[AHumbleOpinion]

Contrary to some of the opinions presented recently it is just fine for Apple to use the 970 and be behind the curve with respect to typical performance. Sure there are specialized apps that can leverage a RISC architecture to outperform x86 or leverage Altivec to outperform SSE, but that is a small minority. Typical performance lags behind PC a little but we are in a situation where PCs and Macs have more performance than most people actually use. Most folks out there in the real world will get along very nicely with a 1GHz PC or a 800MHz Mac. Very few people need 2.xGHz machines, and only a few more have enough disposable income to buy those machines for Quake FPS pissing contests :).

The real Apple problem is that the gap between typical PC and typical Mac performance is starting to grow beyond the range that has historically shown to be viable. Not a problem today, standard dual CPUs counter this to a degree, but it's likely to be a problem in a year or two. While the 970 may only perform like a 3GHz P4 (SPEC), lag whatever Intel/AMD has in a year or two, it will be close enough. Apple will be back to a point where the typical performance gap is small enough. Apple has sold tens of millions of Macs that lagged PC counterparts in performance. They know that their customers are more interested in ease of use. Performance wise close-enough is all they need.

It does with Mac

[Quila]

There's a specific talent to buying a Mac at the right time, as performance increases happen in large steps in a few distinct instances in the year.

PCs just keep getting gradually better and better. But with a Mac you can buy a single processor machine one day only to find you could have had a dual for the same price on the next day.