More on the PowerPC 970

functor writes "Ars Technica's Jon Stokes has a treatise up covering the microarchitecture of the high-performance 64-bit PowerPC 970 microprocessor, due to be released by the end of the year, that goes over in detail how this chip is put together, and how we can expect it to perform. This is the follow-up to Stokes' article detailing the PPC 970's design philosophy. 'It appears to hold quite a bit of promise in bolstering Apple's currently almost obsolescent product line, and it appears to have been designed explictly to fulfil Apple's requirements. To say the least, the second half of this year looks to be pretty interesting as Apple's product line promises to become competitive performance-wise with IA-32 and x86-64-based PCs again.''

Comments

Inaccuracy, Part 1

[11223]

Unfortunately, the vector performance of the G4e has been consistently bottlenecked by Apple's lackluster motherboard and chipset designs--specifically the anemic frontside bus and memory subsystems that Apple has saddled the PowerMac line with.

This implies that the decision of how much bus bandwidth to give the G4e was up to Apple - which it was not. Motorola designed the processor (for Cisco, depending on who you believe), and Apple made do with the anemic MaxBus at 133mhz that they got from Motorola.

Apple'd be putting DDR400 on the G4 right now if they could. None of this (well, except the decision to go Moto) was their fault.

Re:Inaccuracy, Part 1

[Hannibal_Ars]

I should've been more clear on this point. My real problem with the current G4e situation, aside from the 167 SDR FSB, is the fact that it's a shared bus topology, which is just ridiculous. To my knowledge, there's nothing stopping Apple from putting out a chipset that gives each G4e a dedicated FSB (even if it's still 167MHz SDR) to the chipset.

As far as the low MHz and SDR situation, I've also never been totally convinced that Apple wasn't partially to blame for this either, unless they just have zero clout with Moto SPS.

It is competitive !

[Fefe]

Who knows whether it will still be competitive in several months when they actually want to offer it.

On the other hand Apple users won't have much of a choice, and neither has Apple.

Re:It is competitive !

[dasmegabyte]

Sorry, that's bullshit. Apple has whatever choice apple wants to make, and they've had talks with other chipmakers including AMD. If the 970 fell through, they'd go with something else. If they had to, they could switch to an x86 architecture without batting an eye. It worked pretty well for SGI.

Not that it would necessarily be in their best interest. Apple, as a company, represents an ALTERNATIVE, and therefore they try to maintain alternative hardware choices: SCSI over IDE, ATI over NVIDIA, USB over PCI, the one button & metakey paradigm, Flat Panel over CRT -- and the big one, RISC over CISC. Some of these choices have panned out great. Some have flopped miserably...yet despite the doomsayers, Apple is still afloat after 30 years of "forcing" people to buy "crazy" proprietary gear. With "only" 3% of the market, but it's a huge freakin' market, and their margins are gigantic. Part of the reason for this huge margin is that they are the only hardware player, and are therefore able to name their own prices.

As for Apple users having no choices, that's also crap. We always have choices: buy Apple, or buy something else. Keep the current machine and upgrade it, or buy a new one. And Apple will still make the old chips for a long time after the 970 has started smokin' competitors...at least, that's what they did every other time they made a generation jump in processors.

Apple users vote with their dollars and so it's in Apple's interest to do whatever people are most receptive to, which is what they as a company seem to be best at anyway. After all, they got me to spend $538.72 on an mp3 player. This whole "no choice" thing is BS -- you choose the platform, not the hardware. That's the Apple way.

competitive, sure...

[Anonymous Coward]

The PPC 970 will not really make the Macintosh competitive with modern PC's. It will make it competitive with PC's from the beginning of this year, which are not the fastest available any more, and will be even slower when compared to the machines that are available when the PPC 970 ships, which is the very earliest that Apple machines based on it can ship. It will however go a long way to catching up, and take off a lot of the pressure caused by the abominable performance of today's dual processor G4 machines when compared to even inexpensive PC's.

The other unkown in this is the price. PPC 970 based Apple computers may be significantly more expensive. Motorola loses hundreds of millions of dollars each year on their semiconductor business, and IBM does as well. Still, IBM may want to look at Apple and the PPC 970 as a PROFIT center, rather than a LOSS center, like Motorola does with Apple and the G4.

The PPC 970 is great news for Apple, but it is still a bone thrown to them while the x86 PC is feasting on the meat of the Intel and AMD processors.

Dual FPUs!

[EvilTwinSkippy]

Reading through the article, its nice to see some real design going into a processor. Looking through Intel's last few chips, they've been upping ther clock speed and packing in more cache.

Yeah, yeah, they are hog-tied because you can't easily re-compile the entire windows platform to use new instruction sets. Linux users, of course, don't have this problem (muhahahah).

Did anyone else catch the bit on the twin FPU's? I'm just imagining what this thing is going to do with vector operations and frequency transforms.

For most of you non-engineers:

Most 3d vector operations are affine tranformations. Using a 4x4 array of floating point numbers you can translate, rotate, and scale. Works beautifully, but it's a lot of calculations.

The Fast Fourier Transform (FFT) is used a lot in signal processing. It's a floating point monster.

Re:Dual FPUs!

[functor]

Did anyone else catch the bit on the twin FPU's? I'm just imagining what this thing is going to do with vector operations and frequency transforms. For most of you non-engineers:

• Most 3d vector operations are affine tranformations. Using a 4x4 array of floating point numbers you can translate, rotate, and scale. Works beautifully, but it's a lot of calculations.
• The Fast Fourier Transform (FFT) is used a lot in signal processing. It's a floating point monster.

I would imagine that the use of AltiVec/VMX single-instruction multiple-data instructions would be somewhat more effective in doing vector and matrix floating-point computations than scalar floating-point operations as provided by the dual FPUs -- especially on these smaller (4x4, 8x8) matrices. While, in comparison to the MPC 7450, the PPC 970 may have an inferior VMX unit (but with much more bandwidth and a higher clockspeed available), the SIMD instructions allow for more data to be processed per clock cycle than the scalar FPUs can, so I imagine that things like FFTs and vector arithmetic and transformations are better-suited to the use of AltiVec/VMX instructions -- perhaps even hand-tuned to get the best instruction scheduling and highest instructions-per-clock count (and hence computational throughput).

Power 4/PPC970 vs Intel Architecture 64

[tjwhaynes]

With the POWER 4 chips from IBM knocking passed 1.7GHz now, it's a reasonable assumption to make that the PPC970 will clock at similar levels to the POWER 4 chips. So at release, 2GHz plus isn't out of th reach for the fastest chips. Remember that the Athlons with Barton cores aren't clocked much faster than that with the 3200XP clicking around 2.2GHz.

It's not just AMD clocking lower either. The Itanium 2 isn't clocked that fast. Given that 32 POWER 4 1.7GHz processors smoked the 64 Itanium2 1.3GHz processors configuration in the latest TPCC non-clustered benchmark, the POWER and PPC architecture is capable of putting a lot more work through in the same number of clock cycles. There are a lot of nay-sayers trotting out the GHz-is-god line and it is particularly misleading for 64 bit architectures.

Cheers,

Toby Haynes

P.S. Disclaimer - I work in SOFTWARE for IBM, not hardware.

Re:64-bit Adobe apps

[ivan256]

No. And I still don't think that's the reason. How can you make a word processor better using 64 bit code instead of 32 bit code? A spreadsheet? A web browser? An email client? A terminal emulator? A shell? A pdf viewer? I stand by my original point that most consumer apps don't need 64 bit operations.

I anticipated your short sighted response as soon as I hit the submit button. I should have realized that you would think I was talking about optimizing existing applications rather than designing new ones for new problems and I should have said all the rest of this stuff the first time... I even had a horrific vision of the word processor analogy. It was scary. Anyway:

I'm not talking about optimizing existing applications, I'm talking about new applications; programs that do things that we don't use our PCs for now. When we had 8-bit PCs nobody did photo editing or full color page layout on a PC. When we had 16-bit machines nobody used a PC for CAD. Now we have 32-bit machines and are moving to 64-bit. There will be some major tasks that will become possible with 64-bit PCs, but the software isn't there yet because the customers don't have the processors.

Also, I can think of two applications that every single computer user runs that can benefit dramatically from 64-bits, and Microsoft is waiting with the code already written for widespread 64-bit processor deployment to release them: Operating systems and filesystems. Having a 64-bit virtual address space can make your OS much more elegant and efficient since every possible I/O operation can be memory-mapped at once. Similarly, it has already been demonstrated that large relational databases benefit from 64-bit addressing even without taking advantage of the additional memory capacity, and many next generation filesystems will be relational databases.

Sure current consumer apps don't really need 64-bits. If they did we wouldn't have them. It's the PC apps of tomorrow that will benefit. If you don't care to do more with your PC then essentially what you can do today, just more quickly then keep buying 32-bit CPUs. They'll continue to be available for decades...

Performance is out of line?

[Anonymous Coward]

I manage labs of macs, totaling over 300 of them. The users of my labs have no complaints about performance. (They usually only start to whine when the find out the processor speed, but then they are conditioned to think that MHz is MHz; and when asked if their work is slowed, compared to what they do on a PC, they answer "no" it is not slower.)

What is this every non-mac user keeps saying that their performace is out of line with PC's? I have on my desk a hepped up dual P4 and a hepped up dual G4. XP on the dual P4 does not "feel," in day-to-day operations with standard apps like Office or Photoshop, much different from the dual G4. Comparing the MHz does not tell me anything. Using both side-by-side tells me a more "real" story about things. Now, perhaps XP is significantly slowing things down? :-)

I only notice a difference with some high-end 3D apps like Maya or Lightwave, *maybe* also with some high-ed vid apps like Avid's.

I am looking forward to the 970 though. Actually, I am very curious about any 64bit CPU. Hopefully the "growing pains" for anyone to move to 64bit is negligible.