Folding@home in Safe Mode?

[Stonecold]

Would Folding@home run any faster in Safe Mode on Windows? Because there are fewer processes running in Safe Mode, I'd think that that mean less CPU usage, or would being in Safe Mode somehow remove a process or service that enhances Folding@home? I was just thinking this because that I could put my computer in Safe Mode every night before I go to sleep, and if it's faster then it would give me more points - I mean... help find cures faster!

[Jesse_V]

That's an interesting idea. I've never tried that, nor do I remember reading such an experiment. I think that it would run slightly faster since there's less in the way. However, I don't know if it would be able to launch all the necessary drivers, libraries, executables, GUI code, etc, etc, etc. My guess is that that will be your biggest challenge. IF it does actually work, then I'd bet that it'd faster. Safe Mode does shut off a lot of things, but it will be interesting to see how it goes for you.

[Stonecold]

As far as I know, all the necessary drivers are loaded, at least for the command-line SMP client. I'm not sure about the GPU client though. I'd assume it would also be slightly more stable for advanced or beta WUs.

[Leonardo]

I rolled my eyes when I first read the question, then I let it sink in. That's a great question, come to think of it. Why not give it a try? I'd love to try it myself, but my only Windows machine is also my multi-tasking, general purpose machine.

[Stonecold]

I'll try it as soon as I can, but for now I have to send my computer to Geek Squad so they can fix it up (Windows is misbehaving, and nothing I do can talk sense into it). So unfortunately the only computer I'll have for the next weak is a Nintendo 3DS! Not the fastest when it comes to browsing the web...

[PantherX]

...I'd assume it would also be slightly more stable for advanced or beta WUs.

IMO, it won't do much good. The reason is that all Beta WUs (if everything works out fine) will eventually be pushed out to full F@H. That means majority of systems would be running under normal conditions with other applications running and users using the system. Thus this "stable" environment isn't a true representative of a typical donors system. However, if you are in for maximizing the PPD/Science, then that's a different case altogether.

[Stonecold]

Good point... But if I were maximizing for PPD I wouldn't want beta units anyways, because they'd still be really unstable, even in safe mode. And as for science, aren't beta units not so much for finding cures themselves as much as testing to see if the units are ready to be sent to the public to test for cures? So for science I'd guess it would be better to use normal (or maybe advanced) WUs if they're being done in Safe Mode.

[Nathan_P]

Interesting thought, just make sure to enable safe mode with networking - otherwise no net access

[PantherX]

...if I were maximizing for PPD I wouldn't want beta units anyways, because they'd still be really unstable, even in safe mode...

If PPD was your aim, then you would have a lot more options to "choose" WUs (the correct description would be to increase your probability towards a particular WU assignment) as stated by ChasR.

...as for science, aren't beta units not so much for finding cures themselves as much as testing to see if the units are ready to be sent to the public to test for cures?...

Let me put it in another way:
Scientists creates a new experiment (New Beta Project). If each section of the experiment (Beta WU) can be successfully done on the Beta Testers machine (WUs fold successfully), then the scientist will see that the new experiment is performing very well and will later release it to adv and eventually full. If the sections of the experiment fail (WU errors), then it would suggest that something might not be right in the initial experiment so they either try to fix it or it is scrapped.

Hence one can say that folding Beta WUs allows science to progress forward which is also correct.

[orion]

I have a p6993 on w7 running now in safe mode, v6.34

The only data I have on this particular computer close to that is a p6995

TPF look to be similar but it needs to finish so I can get an average.

[orion]

Update:

The p6993 ran 3 seconds per frame faster than the p6995. I'll rerun the p6993 tomorrow in full w7 to see if the TPF actually changes.

Also, the latest installment of v7 will run in safe mode too.

[Stonecold]

The only data I have on this particular computer close to that is a p6995

What you could do (if you don't mind loosing one WU), is run a project in normal mode and allow it to go up to about 99% and then quit the client and delete the data in the work directory. Then boot into Safe Mode and start Folding@home up again. Because you deleted the work data, the core should start from the beginning. That way you can test with not just a similar project, but with the EXACT same WU, which would yield much more accurate results (again, only if you are fine with resetting a WU).

One thing that might be a drawback of using Folding@home in Safe Mode overnight is that that's when system updates, scheduled anti-virus scans, disk defragmenting, recycle bin and temporary files emptying, etc. run because they'd otherwise interfere with your work during the day. So running in Safe Mode might speed up progress, but won't allow some regular, scheduled services, to run.

One thing I've found out is that the Viewer on the v7 beta client actually works (sort of) while in Safe Mode. By "works" I mean it is able to show the demo protein but now adds texture and colors to the individual atoms instead of being a single, untexturized black blob of atoms. Does that mean that the Viewer doesn't display correctly in normal mode because there is some service or driver that's actively interfering with it? Hopefully this piece of information might be able to help fix the problem.

[bruce]

If you delete all of the data in the work directory, the WU will be gone and it may or may not be reassigned. You have to be selective.

Simpler method: A few minutes after a new WU has started processing (and before it reaches the first checkpoint) create a copy of the WORK folder. Later, you can restore that copy, and except for the issue of -forceasm that would apply to a FahCore-78 WU, you should be able to restart the same WU in the same slot as long as the first copy doesn't get to 100%.

There's no reason to proceed to 99%. Run the first 30% twice and let the second copy finish.

[orion]

I was planning on doing something like that. I copied my F@H folder @ 25% on the p6993 and will boot into normal windows tomorrow and replace the present folder with the copied one and get it to start from 0%. It would be the best side by side comparison I can think of...same RCG that way.

In safe mode windows only loads basic video driver at low resolution so that is probably why you're seeing more texture with the viewer.

[Stonecold]

In safe mode windows only loads basic video driver at low resolution so that is probably why you're seeing more texture with the viewer.

I saw no texture in normal mode.