Will be missing deadline on the BIGADV project 2684

[Inik]

I have an i7 @4ghz

Depending on the bigadv, when I'm not using the PC.. the TPF is around 36-37 min. When I'm using the PC (movies or browsing) it's usually around 40 mins.

I have yet to see my 2 gtx 460 fold a 912/925 during a bigadv so I can comment on that situation.

[matsta31]

Back in the days, with linux, the tpf was about 30 minutes. And that was with 2 Gx2 folding. Imade the changes, did not take the gpu out as of yet, Ill see what the TPF will tell me in an hour or 2.

If it matters, I got 6 gigs of ram, runnigng at about 1450 Mhz

[geokilla]

Originally Posted by matsta31

Back in the days, with linux, the tpf was about 30 minutes. And that was with 2 Gx2 folding. Imade the changes, did not take the gpu out as of yet, Ill see what the TPF will tell me in an hour or 2.

If it matters, I got 6 gigs of ram, runnigng at about 1450 Mhz

So the statement I made a few months ago on how a TPF for a -bigadv WU being around 26 minutes was correct!

[LCB001]

Unlike the GPU2 WU's the GPU3's on a Fermi have a significant impact on the CPU, the 48-50 min tpf looks about right for a 2684.

My ppd on a 2684 WU will drop by ~8K+ running two 460's, it's offset by the 28k gain...

[matsta31]

After trying every suggestion, the TPF went down for 1:15 to 1:02... I rolled back all the changes, since my GPU PPD were going down (from 14 000 ish to 9000 ish) . Seems like one core is not enough for the two of them... I ended up killing one GPU client and TPF went back down to 45-48 minutes.

Tonight, I'll try -SMP 6 -Bigadv Flag and managing the CPU affinity so 2 cores are dedicated to the folding.