Folding Forum

I fold on a dual core amd laptop and use the v7 client, I was wondering if the client can still fold while in sleep mode, that way I can shut my laptop lid, and go to to sleep without the fans roaring.

Unfortunately, there is not. If your laptop is asleep then it will use just enough power to be able to wake without having to store the memory contents on the hard drive like hibernation. Also, for laptops if the CPU is going 100% there is pretty much no getting around the fact that the fan will run a lot and probably be noisy too. There are laptop fans and so forth, but generally high-performance computing is best done on desktops, which is a concern for people like us as laptops and mobile devices become more common.

I noticed that the WU is actually making progress very slowly. I changed the power plan to the HP reccomended instead of the high performance power plan, I've noticed no difference in performance and it's much quieter, and I'm folding at around 77C with this power plan, as opposed to 95C with the high performance plan.

If your CPU has a low-performance setting, that's not really the same as sleep mode.

I was folding overnight on the HP-Reccomended power plan. Im using a project 7808 and the TPF was about 1 hour. When I checked the log file just now I saw that the TPF was about 3 and a half hours.

Unfortunately I can't think of a way to know if it was processing slowly or if during three and a half hours, the CPU was sleeping for two and a half or something like that. The TPF is estimated from previous frames and that average could easily be distorted by adding in some time of doing nothing.

I've heard that work units have slow parts. Like a WU that I would normally get a 6 minute TPF would spike to 11-15 minutes then settle back to normal.

Legend2579 wrote:I've heard that work units have slow parts. Like a WU that I would normally get a 6 minute TPF would spike to 11-15 minutes then settle back to normal.

I haven't met a work unit that behaves like that. Usually when TPF spikes, it's caused by background processes like anti-virus updating, scanning, and stuff like that.

iceman1992 wrote:

Legend2579 wrote:I've heard that work units have slow parts. Like a WU that I would normally get a 6 minute TPF would spike to 11-15 minutes then settle back to normal.

I haven't met a work unit that behaves like that. Usually when TPF spikes, it's caused by background processes like anti-virus updating, scanning, and stuff like that.

... or by a power savings setting that reduced the power to the CPU or GPU.

Although in the past there have been cases of "folding events" (a misnomer as the protein may not actually be folding) occurring when lots of the atoms in the simulation come close together. This increases the computation time to calculate position, velocity, energy etc. thus slowing the simulation down. See: http://fahwiki.net/index.php/Folding/Aggregation_Event

uncle_fungus wrote:Although in the past there have been cases of "folding events" (a misnomer as the protein may not actually be folding) occurring when lots of the atoms in the simulation come close together. This increases the computation time to calculate position, velocity, energy etc. thus slowing the simulation down. See: http://fahwiki.net/index.php/Folding/Aggregation_Event

+1 to this. I have dedicated folding boxes with as many as possible background services disabled. I get graphs of the TPF variations using FCI and I have seen random TPF spikes that do not correlate to anything in the event logs. Some frames just take longer.