Phenom F@H Speed?

lelliott731

[H]ard DCOTM April 2017
Joined
Mar 15, 2002
Messages
941
Anyone have a Phenom based system who could give me some numbers for the Windows F@H client? I was thinking about getting a Phenom 9950 2.6GHz Processor with a 790FX Motherboard. Eventually adding a couple of graphics cards.
 
I have a phenom 9750 that was sitting at about 2.8Ghz. Its been a long time since i ran winSMP but i think it was around 1800ppd. Linux SMP with the a2 core was around 2900 i think
 
Ok, thanks for the info sl1982. I guess it'd be better used as a GPU rig then.
 
Ok, thanks for the info sl1982. I guess it'd be better used as a GPU rig then.

I am running the MSI K9A2 Platinum & AMD 9350e to power 4 x ATI 4850.
It has been a very stable foundation for the GPU clients.
I know Tiger has this same Motherboard running nVidia GPUs

Depending on how many GPU's you are going to run on that motherboard:

1 or 2 GPU = I would suggest a AMD 4050e or AMD 4850e
(Both those CPU are very low power draw and run extremely cool)

3 or 4 GPU = I would suggest a quad core AMD CPU
There are so many to choose from. I went with the 9350e due to lower power dray and very low heat output but it is a little more expensive then other AMD quad cores on the market.

I think Tigerbiten has a "Black Edition" cpu in his setup and using some cores for SMP and other cores for GPU if I remember correctly.
 
Ok, according to what numbers sl1982 was getting, it seems the Phenom isn't really that good at F@H if you're running it in SMP mode. Looks like a good cheap GPU platform and that's about it.
 
Its not too bad if you run linux smp. If I run 2 clients i see 4000ppd with the a2 core.
 
AMD in Linux SMP isn't that bad. My two year old 2.8 GHz Opterons (90nm) in a VM produce about 2-3 min slower per frame (A2 WUs) compared to similarly clocked 65nm Intel processors. The newest Phenoms should have comparable performance to a Kentsfield. It's in Windows where the performance gap between Intel and AMD is significant. I'm curious what a 3GHz 45nm Phenom could do.
 
Back
Top