Apple's x86 Move: reality and the party line

a message for Steve Jobs in the contrast between that picture and reality. Steve: either drop the other shoe already or admit you made a big mistake, but,either way, tell the spinmeisters to stop insulting our intelligence

An August 19th, 2005 Computerworld newsletter had an article by Michael de Agonia which caught my eye. Here's a bit from what I thought was an excellent expression of the official Apple line on this:

While the timing of IBM's Power news raised some eyebrows, to me it seems that the announcement was -- at least in part -- a retaliation to Apple's Intel move, a type of PR power play (forgive the pun) to salvage the credibility of the Power architecture.

However, it was a specific detail in IBM's announcement that caught my attention and overshadowed the engineering feat of putting two processor cores on one chip. It was the speed of the chip that surprised me: IBM's new processors top off at 2.5 GHz. That's 200 MHz shy of the high-end chips Apple now ships in its current Power Mac G5s, which offer dual 2.7-GHz G5 processors


As impressive as IBM's dual-core achievement is, Apple is smart to partner up with a chip manufacturer that has shown consistently that it is able to not only push chip technology further, but also to do so within a reasonable amount of time.

Now that's spin - praising IBM for cobbling together a slow but functional dual core 970 and then patting it gently on the back for the achievement while blandly assuming Intel's overwhelming superiority.

In reality Intel is more than a year behind AMD and doesn't have a roadmap that goes beyond trying to close that gap -although that wouldn't matter in terms of Apple's switch to x86 if IBM was even remotely as hapless as de Agonia portrays it to be.

But is it? Well here's a bit from a May 5/05 gameboy interview with Microsoft's Todd Holmdahl (corporate vice president, Xbox product group):

Here for the first time is the entire finalized set of specifications for the Xbox 360:

Custom IBM PowerPC-based CPU:

  • 3 symmetrical cores running at 3.2 GHz each
  • 2 hardware threads per core; 6 hardware threads total
  • 1 VMX-128 vector unit per core; 3 total
  • 1 MB L2 cache


  • 512 MB GDDR3 RAM
  • 700 MHz DDR
  • Unified memory architecture

Memory Bandwidth:

  • 22.4 GB/s memory interface bus bandwidth
  • 256 GB/s memory bandwidth to EDRAM
  • 21.6 GB/s front-side bus

Overall System Floating-Point Performance:

  • 1 TFLOP

That's a three-way, altivec equiped, PowerPC G5 successor, running at 3.2Ghz - available a month before Apple made its announcement and yes, it runs Windows/XP very nicely.

And if that's not enough reality, here's the other shoe: this time from the Sony Playstation spec as given by on May 16/05 and therefore also from the month before Apple's announcement.

PowerPC-base Core @3.2GHz
  • 1 VMX vector unit per core
  • 512KB L2 cache
  • 7 x SPE @3.2GHz
  • 7 x 128b 128 SIMD GPRs
  • 7 x 256KB SRAM for SPE


  • 256MB XDR Main RAM @3.2GHz
  • 256MB GDDR3 VRAM @700MHz

System Bandwidth

  • Main RAM: 25.6GB/s
  • VRAM: 22.4GB/s
  • RSX: 20GB/s (write) + 15GB/s(read)
  • SB: 2.5GB/s (write) + 2.5GB/s (read)

System Floating Point Performance

  • 2 TFLOPS

Ouch, that's another 3.2Ghz G5 successor in there - one integrated with an onboard co-processor GRID to handle lots of processing in parallel.

I've no idea, of course, whether de Agonia is a victim or a collaborator in choosing to present IBM as playing a poor third fiddle to Intel, but there's a message for Steve Jobs in the contrast between that picture and reality. Steve: either drop the other shoe already or admit you made a big mistake, but, either way, tell the spinmeisters to stop insulting our intelligence.