Re: Cell -- WAS: Microsoft edging toward the brink?

["Bryan J. Smith" <b.j.smith@ieee.org>]

On Sat, 2005-02-12 at 15:29 -0600, Robert G. (Doc) Savage wrote: 
> Precisely so.  http://arstechnica.com/articles/paedia/cpu/cell-1.ars

This guy "hit it out of the park" with me off-the-bat:  

  "Ditzel thought that out-of-order execution, register renaming,
   speculation, branch prediction, and other techniques for latency
   hiding and for wringing more instruction-level parallelism out of
   the code stream had increased processors' microarchitectural
   complexity to the point where way too much die real-estate was
   being spent on control functions and too little was being spent
   on actual execution hardware."

Intel had the same view in creating their EPIC/Predication in IA-64.  Of
course, EPIC/Predication flopped and Intel has since been retro-fitting
these traditional run-time optimizations into Itanium2+.  Intel has
_yet_ to retro-fit its 12-year old Pentium Pro core (which is still the
same core of all current Pentium processors) which such (although I
predict the, still fictional I know, "Yamhill2" will).  The Digital
Alpha team predicted this when Itanium was first announced, and AMD
adopted these approaches from Alpha in their 1999 Athlon core (based on
earlier work done in their 1993 NexGen core).

My take on both IA-64 and Crusoe (WARNING:  flammable ;-)
It's programmers attempting to trying to solve the issues of
semiconductor design that is best left to engineers.  Hence why and how
companies like IBM microelectronics knows what they are doing.  I know
that seems arrogant, but CS math-based theories are _never_ applicable
to EE materials, digital and analog electronic implementations.

But regardless of semiconductor design theories, it's nice to have an OS
that has an API not fixated on a single architecture.


-- 
Bryan J. Smith                                  b.j.smith@ieee.org 
---------------------------------------------------------------- 
Community software is all about choice, choice of technology.
Unfortunately, too many Linux advocates port over the so-called
"choice" from the commercial software world, brand name marketing.
The result is false assumptions, failure to focus on the real
technical similarities, but loyalty to blind vendor alignments.



-
To unsubscribe, send email to majordomo@silug.org with
"unsubscribe silug-discuss" in the body.