Pentium D and Prime95

[nomadicus]

Originally Posted by Joshua2

Which if your running two p95's should be just about twice as fast . . .

Should be about 1.7x to 1.8x as fast. I have a dual Xeon 2.8GHz, both CPUs running LL with affinity set, it is getting 1.6x performance.
Hopefully, someone will post the results of prime95 on a dual core in the perpetual benchmark thread.

[georgekh]

when i said that they are out and i want to buy one i meant that in a few months i will get one because i know they are still not available here where i am and i dont mind if the chips are hot because i got a Vapochill so thats no problem :)

[dsouza123]

In my previous post listing dual core Athlon 64 and Pentium D/EE the L2 cache
size for the Athlon 64s were half of the correct value
(I mixed per core L2 cache for AMD with total L2 cache for INTEL)

It should have the following instead.

==========================================
All L2 cache values listed below are total cache.

For the Athlon 64 X2
Clock, Rating, L2 Cache, FSB
----------------Athlon 64 X2 (Manchester Socket 939)
2200 4200+ 1MB 400
2400 4600+ 1MB 400
----------------Athlon 64 X2 (Toledo Socket 939)
2200 4400+ 2MB 400
2400 4800+ 2MB 400
==========================================

[JHagerson]

Here is what I have found. To get Prime95 to use both cores of the Pentium D, the affinity for one instance needs to be set to processor 0, the other instance to processor 2.

EDIT: I take this back. I can get both instances on one core or the other. George, please tell me what I need to do! Thank you.

[delta_t]

I think you want affinity as CPU 0 and 1 (not 0 and 2).

[Cruelty]

Originally Posted by JHagerson

Here is what I have found. To get Prime95 to use both cores of the Pentium D, the affinity for one instance needs to be set to processor 0, the other instance to processor 2.

If this is a Pentium D you have and not the EE (the later being HT enabled thus providing 2 physical and 2 virtual CPUs) you should only choose between CPU0 and CPU1 when setting affinity, as suggested by delta_t above.