New Cell Article, on research.ibm

Carl B

Friends call me xbd
Legend
This article was brought to my attention earlier today, and it's a good read. Basically a complete summary - from IBM's side - on the Cell architecture: it's conceptualization, design, and implementation.

The initial 'history' section starts off a little slow, especially after reading some of the other 'histories' out there for this chip (where's Kutaragi's 'aesthetics' for example? Or IBM demanding Power be included?).

It quickly picks up pace however going into the other sections.

Some sections may stand out more than others, but the whole of the article is interesting.

http://www.research.ibm.com/journal/rd/494/kahle.html

Here's a pipeline breakdown as well from the article for the PPE:

kahle2.gif
 
Lysander said:

Well read it! :p

PS - Thanks for the honor Acert. :cool: Original credit must go to Rukawa at PSINext for bringing it to my attention. Really good read though - I love reading about why engineers determine to go with a certain design aspect as opposed to an alternative - and this article deals with these things even down to the chip packaging.

I mean, this is the kind of goodness you can expect to find:

...In the case of the Cell processor, having eliminated most of the barriers that cause inefficiency in high-frequency designs, the initial design objective was a cycle time no more than that of ten fan-out-of-four inverters (10 FO4). This was later adjusted to 11 FO4 when it became clear that removing that last FO4 would incur a substantial area and power penalty...
 
Last edited by a moderator:
Ah well, I was wondering if it'd been posted anywhere else in the forum; must've missed that one somehow. Still, it really is a good article, so even outside the context of PPE and it's VMX configuration, it's certainly worth the read.

Certainly for myself the VMX aspect of the article was only one of a number of interesting article facets.
 
Back
Top