Intel ACK!

Article from theregister :
Second-class Intel to trail AMD for years

http://www.theregister.co.uk/2005/10/29/intel_xeon_2009/print.html

"Underneath all of these roadmap adjustments lurk some painful technology slips that must have customers concerned. In particular, it now appears that Intel will stay married to its front side bus dependency for much longer than previously expected and will fail to deliver integrated memory controllers on time. Where Intel had a very real shot at closing the gap with AMD in just 18 months on previous roadmaps, it now looks more likely to trail for close to four years. This should worry folks at Dell, HP and SGI, as they're most vulnerable to Intel's shortcomings."


I think Steve Jobs is a creative genious, but he absolutely blows on his timing for moving to a new processor =(

On a brighter note, Intel makes the best laptop chipsets and CPUs and that's not expected to change much, so at least Apple will be able to make some really high end laptops.

depends on how the MacIntel chips are designed / fabricated.

Are there details on the chip arcitecture that the new macs are going to use?

Could it be possible that steve has more than we know comming? or maybe the macintels never do come otu since the unplanned dual core is out?

The Intel-powered Macs will come out. The laptop processors are the REASON they switched, not the desktops. If they thought they could've produced mobile G5s, they wouldn't have even bothered to go to Intel.

The Register is known for it's (often legitimate) criticism of Intel. Anything can happen in 18 months. I'm going to worry about this when it happens. Meanwhile, the PowerPC Macs are fairly competitive, for the time being.

im not worries TBH intel have a good while to imporve the P4 before apple start to think about releasing an intel PM

Putting things in perspective (as an ex-Intel guy) the performance deltas between AMD and Intel are always fairly small - so unless you run benchmarks for kicks or to impress people the actual user differance is small.

Not that Intel doesn't screw-up - I've seen plenty of lost opportunities and goofs. I can say from my last meetings on future x86 CPU architecture - they are aggressively moving into multi-core - and multi "virtual machines" via Vanderpool technology. Also as I've mentioned before - Intel is leading the way on the "serialization" of the PC ...... 64bit or 128bit buses are just too hard to route (economically) and maintain sub-nano second timings. So all buses will move to groups or clusters of very high speed serial links over time. The new PCI-Express is a good example of this.

So - they have missed and messed up some product (cpu/chip) launches - but the company is really going thru a MAJOR direction and architecture change ---- a lot of this included internal battles with the "old guard" just wanting to use the tried and true brute force method of cranking up raw clock speed - rather than the more complex - elegant - and effective of approach of more CPUs running at slightly slower speeds and improving overall system performance via these new serial buses

Ex_PC_Puke said:

Putting things in perspective (as an ex-Intel guy) the performance deltas between AMD and Intel are always fairly small - so unless you run benchmarks for kicks or to impress people the actual user differance is small.

Not that Intel doesn't screw-up - I've seen plenty of lost opportunities and goofs. I can say from my last meetings on future x86 CPU architecture - they are aggressively moving into multi-core - and multi "virtual machines" via Vanderpool technology. Also as I've mentioned before - Intel is leading the way on the "serialization" of the PC ...... 64bit or 128bit buses are just too hard to route (economically) and maintain sub-nano second timings. So all buses will move to groups or clusters of very high speed serial links over time. The new PCI-Express is a good example of this.

So - they have missed and messed up some product (cpu/chip) launches - but the company is really going thru a MAJOR direction and architecture change ---- a lot of this included internal battles with the "old guard" just wanting to use the tried and true brute force method of cranking up raw clock speed - rather than the more complex - elegant - and effective of approach of more CPUs running at slightly slower speeds and improving overall system performance via these new serial buses

Click to expand...

I would agree that Intel lacks behind AMD slightly in performance. However, the small difference in performance is compounded a vast price difference.

It made me sad to see Apple choose Intel over AMD. I am always rooting for the underdog, especially when the underdog has a betterr product at a lower price.

Well it gets worse. I never liked the switch idea so maybe I'm bias, but check this out :

P.A. Semi's major PowerPC announcement, and looking back at The Switch

http://arstechnica.com/news.ars/post/20051026-5486.html

Check out these stats, distilled from the press release:

* Two 64-bit, superscalar, out-of-order PowerPC processor cores with Altivec/VMX
* Two DDR2 memory controllers (one per core!)
* 2MB shared L2 cache
* I/O unit that has support for: eight PCIe controllers, two 10 Gigabit Ethernet controllers, four Gigabit Ethernet controllers
* 65nm process
* 5-13 watts typical @ 2GHz, depending on the application

Yup, they appear to have crammed all that stuff onto one laptop-capable chip, and it's due out sometime in very late 2006 or early 2007.

...

P.A. Semi., IBM's 970FX/MP, POWER5, and the next-generation of game console chips all show that the book is by no means closed on PowerPC. In fact, we're in the opening moments of a full-blown PowerPC Renaissance. Who knows how long the PPC revival will last, but Apple would've been hard pressed to pick a worse time to jump ship for performance or performance/watt reasons.

Click to expand...

So maybe I was right all along...

lonerider said:

Well it gets worse. I never liked the switch idea so maybe I'm bias, but check this out :

P.A. Semi's major PowerPC announcement, and looking back at The Switch

http://arstechnica.com/news.ars/post/20051026-5486.html

Click to expand...

When the Intel switch was first announced in early summer, I said "hey, maybe Apple is announcing this to light a fire under IBM's butt to produce better chips more quickly!"

But nooooo... everybody quickly crushed my comment, saying that "that's not how it works."

Well, now we have dual-core Powermacs... Furthermore, according to all that I've just read, Apple announced a switch to Intel, and a few months later... WHAMMO! Ppl are saying their timing sucks, that there's lots of life left in PPC. And there appears to be technical data to back that statement (see the above post.)

MY NEW THEORY: Maybe they'll use Intel for all the mobile stuff, but keep a line of PPC-driven Powermacs open for those of us who remain "old school." I'm not sure how that'll work from a SOFTWARE point-of-view, but it WOULD be a way to have the best of both worlds.

Anybody care to comment?

sevenhelmet said:

Anybody care to comment?

Click to expand...

Only to say .. Told you so!

Everyone thought I was daft a few months ago but it looks like I hit the nail on the head.

I was even told that the AMD 64 chips rule so I went out and built myself a kick-arse AMD 64 based system, my 2 year old twin G5 runs rings around it, personally I'll be buying a quad G5 in the new year.

Amen-Moses

Unfortunately, the POV that the PowerMacs and professional range of machines will be the first with Intel x86 is an endemic opinion even in Apple retail channels...

The truth of the matter is, what were the first PowerPC Macs that moved on up from the then current MC68030/68030 CPUs? The Performa.

Consumer Macintoshes.

If anything, I don't expect to see a intel PowerMac until the end of 2006 at the earliest. The Universal Binary architecture is very good, and makes much more sense than a pure emulation based system, as per the original PPC 601 Macintoshes emulating the 68000 instruction set with no dual-binary format, or indeed Apple opting to rely on emulation on the intel purely through Rosetta (PPC Emulation).

Through all the marketing and ADC related gumph that describes Rosetta as a very fast PPC emulator, the fact of the matter remains: I doubt we'll see fast G4 with Altivec emulation on the first Mac-intels. I realistically expect low-end PPC 750 speeds (G3).

Essentially we end up with the reverse situation we have at present with VirtualPC, except the Intel Pentium architecture much like the G5 does not do byte order swapping as a built into the CPU. On the Mac on PPC, we have big-endian systems, in line with the 68000 architecture, the most significant bit (MSB) is on the right so the Mac reads right to left. The intel platform is little endian and the MSB on the left hence reading left to right. In order to get somewhere near code translation, the MSB has to be switched; much as Rosetta will have to in order to emulate a PPC processor, and thus as we all know - VirtualPC which has to do a similar thing hardly emulates a fast PC on a fast G5 system.

What's the reason for me talking about this?

It's just more evidence that until applications are fully optimised for intel, or indeed available for intel, and a processor that is significantly more powerful than the G5 comes from that stable - the pro-end will not switch. After all, why would anyone switch to an Intel PowerMac if it were to offer bad performance over its PPC cousin.

This all ratifies what I have said all along but has been rubbished very often - Apple will wait for the intel Mac to mature a little before such an intel PowerMac, as in doing so early would be risking mission critical applications and systems.

Vicky

Amen-Moses said:

Only to say .. Told you so!

Everyone thought I was daft a few months ago but it looks like I hit the nail on the head.

I was even told that the AMD 64 chips rule so I went out and built myself a kick-arse AMD 64 based system, my 2 year old twin G5 runs rings around it, personally I'll be buying a quad G5 in the new year.

Amen-Moses

Click to expand...

Vindication!!!! I'm a PowerPC guy all the way- that was a selling point for me when I bought my mac.

I think that Intel is a better x86 partner for Apple at the moment. (Grumble mumble PowerPC just fine mumble)

AMD is a processor builder. They build high-performance x86 CPUs better than anyone else. But that's all they do.

Intel is a larger company, which does research into "platforms"--CPUs, architectures, chipsets, compilers, and other related technologies. If you're just building a faster PC, all that extra stuff just gets in the way and drives up costs. But if you're a major hardware vendor in the process of a major transition, that extra knowledge can be very helpful, especially when it comes to optimization.

Once Apple has moved successfully to x86, then AMD will be more viable. Until then, maintaining a friendship with Intel may bear fruit in matters other than straight speed.

lil said:

The truth of the matter is, what were the first PowerPC Macs that moved on up from the then current MC68030/68030 CPUs? The Performa.

Consumer Macintoshes.

Click to expand...

I remember differently. Maybe things were different on the other side of the pond, but I remember the Power Mac 6/7/8100's being introduced.

The first Performa to go PowerPC here was the Perforama 6115, which was really a Power Mac 6100.