There is a lively discussion going on on Anandtech forums about how Bulldozer will be a good folder, with some interesting tidbits from JF-AMD. It get a bit technical for me so I was hoping someone could break it down for me. Is F@H a floating point hogger? How does this module's INT and FP capabilities compare to one core from Sandy-Bridge? Most important, will it fold well?
Thanks beforehand.
Excerpt from page 4:
http://forums.anandtech.com/showthread.php?t=2148217
Excerpt from page 4:
We have a 256b FP datapath (pipes 0 and 1) AND a 256b INT datapath (pipes 2 and 3), so
2 128b FP + 2 128b INT
or
1 256b FP + 2 128b INT
or
1 256b FP + 1 256b INT
or
2 128b FP + 1 256b INT
The INT here is an integer unit for doing the integer portion of math inside an SSE instruction, that is not the integer clusters that you would commonly call cores.
Plus there is a really cool feature around moves. Technically, we can do 4 128b SSE moves per cycle with a ZERO cycle latency. This is known as MOVE ELIMINATION.
http://forums.anandtech.com/showthread.php?t=2148217