Reverse Hyperthreading is what would make the most sense in this type of configuration. According the the blog this would be made for gaming enthusiast......
Reverse Hyperthreading 2 cores attached to other 2 via hypertransport bus would POWN ALL !
Well If this were to be true.
2x 3800x2 as a kit for 600 bucks would be great:
2 Processors on 1 Die with Reverse Hyperthreading to give 2 7600+ x2 connected with hpertransport...
Extra bandwidht from dd2 will be utilized to the fullest....
Reverse Hyperthreading is what would make the most sense in this type of configuration. According the the blog this would be made for gaming enthusiast......
Reverse Hyperthreading 2 cores attached to other 2 via hypertransport bus would POWN ALL !
Well If this were to be true.
2x 3800x2 as a kit for 600 bucks would be great:
2 Processors on 1 Die with Reverse Hyperthreading to give 2 7600+ x2 connected with hpertransport...
Extra bandwidht from dd2 will be utilized to the fullest....
ANY THOUGHTS :-)
MMM do you ever get tired of making new sock puppets?
My take on this : 4x4 platform is a good idea . IT brings us back to the old days of cheap dual socket boards (P3, Celeron, Athlon MP), allowing you flexibility (buy one cpu now, add another later) or affordable multi socket solutions (workstations and the like).
But its a horrible idea if you think of it as a counter to Conroe,especially for gamers, because for most people, 2 cores/threads is already overkill. Precious few will benefit from 4 thread capability for the coming years, but for some (3D renderers, video encoding,..) it could be an excellent platform. ironically, in the same way P4 was once against Athlon XP.
As for the reverse hyperthreading.. I don't see it happening. AFAICT, only horribly written code would benefit from being magicaly, automatically split into different threads and being exectued on different cpu's that are hundreds if not thousands of clockcycles "away" from each other. Properly written multi threaded code would likely be orders of magnitude faster, and I suspect for almost other code, running it on a single core would be faster too.
You are about to answer a thread that has been inactive for more than 6 months. If you still wish to proceed, please ensure that your posting is original and does not duplicate or overlap any prior responses to this thread.