Letting cuda WUs use a full thread/core

Joined: 4 Jun 08
Posts: 35
Credit: 88,264,743
RAC: 0

RE: RE: My GPU runs 4 of

My GPU runs 4 of those at a time completing them (if I let it crunch without me using the PC) in about 92 minutes. This means my GPU could complete about 62 BRPcuda tasks in 24 hours. Of course I do use my PC so it can't quite reach that numer but still.

well - this looks like it works exactly for your configuration. the i7-2600 seems to be fast enough to do this. things might even change if you run other apps as CPU-tasks. HT-machines usually improve if you mix 2 or 3 apps with different demands.

anyway it would be interesting to see what happens if you disable HT via bios..

Yes I tried that but not for long. I'll have to do some tests to see if it helps complete the tasks faster with ht off. What I did notice immediately is that any running CPU tasks seemed to cause a lot of trouble and starved the GPU.

Weather forecast says there'll be some bad thunderstorms coming my way soon, if those are gone by the time I go to sleep I'll let it run for the night with ht off and no running CPU tasks.

No I haven't signed up to that mailing list, perhaps I will.

PS I have an i7 870, that i7 2600 I only used as an example of how long a CPU BRP task runs compared to my BRPcuda tasks, I just looked at a few of my validated results to find one which got paired to a CPU wu run by a fast CPU.

PPS I'm really glad I got my CPU cooler installed last week, it's about 32c outside right now and my CPU is stable at 60c. With the stock cooler at this clockspeed Im sure it'd be well on it's way to the thermal limit of 105c (I think it was 105, could be 115 but way too hot anyway).

Joined: 31 May 05
Posts: 53
Credit: 627,005,014
RAC: 0

RE: RE: I did a few


I did a few experiments with CPU priority of BRP3 (changing the nice value for BRP3), bud did not see any noticable improvement at all...

Ubuntu 10.04, GTS250, Athlon2 4core, gpu driver 270.29.
4 CPU units and 2 BRP3 units at the time...

this is not enough - limit boinc to using 75% of your CPU's and watch GPU-tasks improving on speed.

Well, here they are some results:
CPU 4 cores, 4xCPU units at time, 2x cuda units with low priority
GPU 2 units at time, ---> 2h 42min (+ - a few minutes)

CPU with only 3units at time, 2x cuda units with a real-time priority
GPU 2 units at time, ---> 2h 32min

It means improvement (6-7)% on GPU, but only 3 CPU units at time instead of 4 ones
From my point of view - nothing much...

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.