pari
07-06-2006, 12:16 AM
Its by IBM employees so it has both CELL and 360 related articles
http://gametomorrow.com/blog/index.php/category/cell/
And from that site about CELL
http://gametomorrow.com/blog/index.php/2006/03/24/cell-cant-texture/
Much has been said about Cell’s presumed inability to texture map well. Given the small (256KB) local stores and DMA memory access, the SPEs were relegated by many to only handle nice streaming geometry type workloads. This seemed like an issue ripe for a little prototyping.
First, colleague Mark Nutter, implemented a software cache abstraction layer for the SPE giving us the ability to both hide the complexity of DMAs and benefit from transparent data reuse. Next, given the lessons learned from this paper, we tiled our textures, optimized our access patterns, and implemented several cache replacement policies. We then rewrote the shader in the Quaternion Julia Set Raytracer to add five cubemap texture lookup passes - 3 refraction lookups, a reflection lookup, plus a background lookup. These five texture lookups were then blended together with a fresnel calculation and modulated with the base lighting computation to form the final sample color.
The results were very pleasing.....
To do a relative comparison regarding production, from the same site about XeCPU
http://gametomorrow.com/blog/index.php/2005/10/27/xbox-360-cpu-details-described-at-mpr-fall-processor-forum
We developed an XBOX 360 unique CPU chip, engineered and optimized for the specific product constraints. We went from 1st silicon to volume production in 8 months and if you get in line early enough you might be able to buy one November 22.
So compared to this CELL was taped out in Jan'04 (from ISSCC'05 presentation), I personally reviewed the CELL paper for that conference, and ISSCC requires to have Silicon data for the presentation. So by end of '04 CELL Silicon data was available. Now IBM/SONY had about two years to get CELL debugged for production in 90nm, so I doubt CELL would be shortage. This two years from Jan'04 onwards, it given IBM would have started 65nm remap work, so by middle of '05 it would have been taped out or by the latest at end of '05. So from begining of '06 CELL 65nm production debug would have started, Intel started shipping 65nm from about begining of this year and IBM is not far behind.
IMHO, I strongly believe PS3 would have CELL in 65nm but I can be wrong, dunno what is the state of IBM fab. Since Logic bugs would have been squashed in 90nm, so 65nm revs would be production issues.
http://gametomorrow.com/blog/index.php/category/cell/
And from that site about CELL
http://gametomorrow.com/blog/index.php/2006/03/24/cell-cant-texture/
Much has been said about Cell’s presumed inability to texture map well. Given the small (256KB) local stores and DMA memory access, the SPEs were relegated by many to only handle nice streaming geometry type workloads. This seemed like an issue ripe for a little prototyping.
First, colleague Mark Nutter, implemented a software cache abstraction layer for the SPE giving us the ability to both hide the complexity of DMAs and benefit from transparent data reuse. Next, given the lessons learned from this paper, we tiled our textures, optimized our access patterns, and implemented several cache replacement policies. We then rewrote the shader in the Quaternion Julia Set Raytracer to add five cubemap texture lookup passes - 3 refraction lookups, a reflection lookup, plus a background lookup. These five texture lookups were then blended together with a fresnel calculation and modulated with the base lighting computation to form the final sample color.
The results were very pleasing.....
To do a relative comparison regarding production, from the same site about XeCPU
http://gametomorrow.com/blog/index.php/2005/10/27/xbox-360-cpu-details-described-at-mpr-fall-processor-forum
We developed an XBOX 360 unique CPU chip, engineered and optimized for the specific product constraints. We went from 1st silicon to volume production in 8 months and if you get in line early enough you might be able to buy one November 22.
So compared to this CELL was taped out in Jan'04 (from ISSCC'05 presentation), I personally reviewed the CELL paper for that conference, and ISSCC requires to have Silicon data for the presentation. So by end of '04 CELL Silicon data was available. Now IBM/SONY had about two years to get CELL debugged for production in 90nm, so I doubt CELL would be shortage. This two years from Jan'04 onwards, it given IBM would have started 65nm remap work, so by middle of '05 it would have been taped out or by the latest at end of '05. So from begining of '06 CELL 65nm production debug would have started, Intel started shipping 65nm from about begining of this year and IBM is not far behind.
IMHO, I strongly believe PS3 would have CELL in 65nm but I can be wrong, dunno what is the state of IBM fab. Since Logic bugs would have been squashed in 90nm, so 65nm revs would be production issues.