work units not sending. Also, idle not working properly

[bruce]

Note that you have four slots. You may not realize that they can be controlled individually.

While may have intentionally chosen to have them all pause until the system is idle, that would not have been my choice.

I see no reason to pause the CPU slot. Period. It will not interfere with whatever else you do with your system (unless you have extremely small RAM, such as 512K).
I do understand why people choose to pause the GPU which drives their screen. A system can feel "sluggish" if the GPU is busy folding and mouse clicks are delayed by a perceptible amount.
Pausing GPUs which are NOT connected to your primary display are much more problematic.

No matter what you choose, your should understand your options and make a conscious decisions on how you want to balance what FAH uses with what resources you wish to dedicate to other programs..

[Volnaiskra]

Looking in C:\ProgramData, C:\Program files, and C:\Program files (x86), I see only one FAH folder. Does that indicate that I only have one version installed, or is there another way I can check?

Is it possible that it's somehow running two versions on startup? I can see a FAH entry in msconfig AND in Start Menu\All Programs\Startup. Could that be a problem?

Whenever I get this greyed out problem, Web Control always times out (neither of your links worked).

[PantherX]

Assuming that you used default locations, if you installed for all users, then:
C:\ProgramData\FAHClient
C:\Program files (x86)\FAHClient

If you installed for yourself only, then:
%AppData%\FAHClient
C:\Program files (x86)\FAHClient

Not sure about the start-up since I don't use automatic start-up, only manual. A system search for the folder FAHClient should only show two (or three if you include the Start Menu) results in a normal installation.

If you have installed correctly, what I would suggest that you do is to disconnect the LAN cable (or disable the NIC) and move the config.xml to a different location and then start-up FAHClient. This should force it to generate a new default config.xml and so FAHControl should be able to connect to FAHClient (assuming that no security application is blocking the connection). The reason for disconnecting the LAN is to avoid downloading any WUs while you are testing the set-up. Do note that without internet connection, Web Control may not function but FAHControl should fully function. Also, is there a pattern to when the greying out occurs, i.e. after every X minutes, after Y WUs completed, etc?

[Volnaiskra]

I've only got the two locations, so that seems right.

I deleted config.xml and restarted. FAH is still greyed out, though it is folding - CPU and all GPUs are churning away. FAHControl has seemingly no control over FAH. (I can right-click it in the system tray and change from "medium" to "pause" or "idle", but it doesn't make a difference, other than putting "zzz" on the icon when I make it "idle").

I deleted the FAHclient shortcut from Startup, and restarted PC. FAH didn't start up - so that quells any suspicions about it being accidentally started twice when I turn the computer on. I then turned it on manually, with the same results as before (folding but greyed out).

I haven't noticed any pattern with the greying out, other than it doesn't seem to happen after a fresh install of FAH. When I reinstall FAH (and I've done this many times since getting this problem), things seem to work. Then, some of the cores start not getting closure acknowledgements. Then, maybe a week or two later, FAH starts greying out and becoming unresponsive.

I haven't

[bruce]

Greying out can happen if the shortcuts don't all point to the same "Start in" directory or if that directory happens to prohibit FAH from writing there. Greyed out can also happen if you didn't happen to notice that your firewall blocked some of the features of one of the FAH programs.

If you haven't tried it since your last FAH install, disable your firewall and see if starting FAHControl gives you an online status. At least that will demonstrate that you need to adjust the firewall settings.

[Volnaiskra]

Sigh. Not sure what's going on or where or why.

I disabled the windows 7 firewall, then loaded up FAH, and it worked. But I've since enabled the firewall again, and FAH still works. So I don't know if the firewall was the problem, or if it was just a coincidence.

Anyway, folding is working now, but I'm still getting the missing acknowledgements (ie. upload gets to 99% and then I never hear from it again)

[Volnaiskra]

Back to the same motley bunch of problems: FAH ignoring idle setting and folding even while idle and even when I try to manually pause slots, completion acknowledgements not working, GPU folding but CPU not folding, client periodically greying out.

I've tried all the advice given here so far, but the problems still recur. FAH is really spazzing out, and I don't think the problem is deeper than clogged bandwidth or a bad setting. I wonder if there's some sort of hardware or software conflict happening. I'll try running it for a while with all background programs closed and see if that helps anything.

Not sure if this is relevant, but FAH did a really poor job of identifying my GPUs once I bought a third card for PhysX about 6 months ago. Back when I only had two Titans, it was all good. But once I introduced a gtx 650, FAH got really confused about what card was what: it would frequently refer to the 650 as a Titan, and one of the Titans as a 650. It did this whether I manually set the slots or just left them all at -1.

I've since swapped my 650 for a 650ti, and reinstalled FAH. FAH now seems to the 3 cards properly. However, my previous experience with the 650 showed that FAH can't always be trusted to know what's what. Perhaps it's still somehow getting confused about my 3 GPUs now.

[bruce]

Are you absolutely sure you don't have two independent clients running?

If you add or alter a GPU without a clean install (no config, no slots, no WU) GPUs are likely to be misidentified. You CANNOT make a small change to your GPU configurations without the possibility of problems. It always seems to sort itself out with a clean install.

Post the log showing where you paused a GPU.

[Volnaiskra]

OK, well at least I know that the misnamed GPU thing is a known issue. It's been fine for months though, since I did one of my many clean installs

You know, I have a hunch that it IS running two clients, but I don't know what I can do about this, or even how I can confirm it. I've only installed it once, and I've triple checked this. Yet something tells me it still manages to force two versions of itself into memory. I no longer use the auto-startup method, hoping to avoid this, but that hasn't helped.

I ran it with all background programs closed, but that didn't help. I have just now done a clean uninstall again (so I can't post that log, sorry). During the uninstall, the uninstaller paused and said "shutting down active clients", and froze indefinitely.....this seems to be another incarnation of the same fundamental problem I'm getting: FAH just doesn't seem to have proper control of its own processes on this PC.

As I manually shut down the processes, the uninstaller continued. There were about 4-5 processes to shut down. When I shut down 2 or 3 of them, the uninstaller started continuing, then gave me a popup saying "can't continue because FAHClient is running". So, it first *thought* that FAHClient was turned off, but then realised that it wasn't. I dunno - that also kind of sounds like there are two versions running, or that FAH isn't tracking the entire operation properly.

Anyway, once I manually closed all the proccesses in task manager, the uninstall successfully completed. I can't find any trace of FAH on my PC, and I will now reinstall.

[bruce]

By default, FAHClient will be installed as a background service. The installation guide cautions against starting a second copy manually. If you choose to pause processing, you can manage that through either Control program and although FAHClient will remain running, it will be idle until you restart one or more slots.

[Volnaiskra]

I haven't been starting second copies manually.

Originally, I was running it the default way (FAH runs at startup). Then, based on PantherX's comments, I decided to uninstall and try installing it without the auto-startup feature. So now, I run it (once) manually to start it. Either way, it didn't make any difference, since I was getting the same sorts of problems in both scenarios.

I know I can pause processing through web control or advanced control. The problem is that sometimes FAH shits itself and stops being able to control its processes. I pause it, but FAH continues to take up 99% of my GPU and/or CPU. Or, I set it to "idle", and the FAH icon gets zzz's on it, yet it continues to fold furiously even while I'm using the computer (this last problem is referenced in the original thread title).

Look, I think if we're going to solve this, we need to look past basic user error. I'm prepared to accept that it may be PEBKAC, but only of some more complicated and/or leftfield variety. I think the basic stuff has been covered already. I've tried all the suggestions you guys have given me, but the problems keep recurring. So, I don't think it's something basic like bandwidth problems, firewall settings, or me using the software the incorrect way. I think it's more likely to be some sort of hardware or software conflict, bug, or else I've gone and screwed something up in some creative way.

The thread is long now, so I'll reiterate my problems here:


-Folding usually works fine
-But some WUs fail to receive a completion acknowledgement. That is, they complete, they upload, I get points, but they still say "sending" on my FAHcontrol, and never go away. After they get to about "upload 99%" there is no more trace of them in the logs
-Eventually, these 'unsent' WUs pile up.
-Eventually (not sure if this is related to previous problem or not), FAH starts becoming less reponsive
-Sometimes FAHcontrol is inaccessible (web control times out, advanced control greys out) even though folding itself continues
-Sometimes FAHcontrol loses the ability to properly control the cores: pausing it doesn't pause them, they continue folding even when I set it to "idle", and the uninstaller isn't able to successfully close them down
-It seems that after about a couple of weeks, these problems start occurring, and get progressively worse
-After a clean install, the same pattern restarts: works fine, then slowly the above problems start coming. I've done about a dozen clean installs this year, using various different settings (default, custom, autostart, manual, beta, normal), with no difference.
-In the past, I folded just fine with this PC (same windows installation, mostly the same hardware), and didn't have any of these problems. The only changes I can think of between now and then are a new PhysX card (GTX650ti), and a new netgear router. The router, and my internet in general, works fine in every other department (browsing, steam, streaming, downloads, home network, streaming PC games to nvidia Shield). I've also started running nvidia Experience since that time.

My system:

Code: Select all

07:44:41:******************************* System ********************************
07:44:41:          CPU: Intel(R) Core(TM) i7-4770K CPU @ 3.50GHz
07:44:41:       CPU ID: GenuineIntel Family 6 Model 60 Stepping 3
07:44:41:         CPUs: 8
07:44:41:       Memory: 15.93GiB
07:44:41:  Free Memory: 12.68GiB
07:44:41:      Threads: WINDOWS_THREADS
07:44:41:   OS Version: 6.1
07:44:41:  Has Battery: false
07:44:41:   On Battery: false
07:44:41:   UTC Offset: 10
07:44:41:          PID: 8836
07:44:41:          CWD: C:/Users/Dave/AppData/Roaming/FAHClient
07:44:41:           OS: Windows 7 Home Premium
07:44:41:      OS Arch: AMD64
07:44:41:         GPUs: 3
07:44:41:        GPU 0: NVIDIA:3 GK110 [GeForce GTX Titan]
07:44:41:        GPU 1: NVIDIA:3 GK110 [GeForce GTX Titan]
07:44:41:        GPU 2: NVIDIA:3 GK106 [GeForce GTX 650 Ti]
07:44:41:         CUDA: 3.5
07:44:41:  CUDA Driver: 6000
07:44:41:Win32 Service: false

Other points of note:
-I'm running Win7 64bit Home Edition
-I've seen very few folding errors in any of the logs I've looked at
-My CPU is overclocked. It's stable with overnight stress tests and regular usage, with temps well within the safe limits
-My motherboard is an Asrock Z87 Extreme9/ac. Overall it works fine, but it seems to have some minor problems with USB not always getting power at startup (sometimes can't enter BIOS because of this - have to reset PC couple of times until USB gets power).
-My Titans use a custom vbios by svl7 that removes nvidia's GPUBoost2.0 feature. In other words, it gives the Titans stable clocks rather than variable ones. I also have them mildly overclocked. They work without problems in heavy gaming, with safe temps.
-The Titans are set to SLI, and the 650ti is dedicated to PhysX

[7im]

After pausing, how long did you wait for the usage to drop? CPU folding usually stops in a few seconds, while GPUs can take up to a minute to stop.

[bruce]

Are you running with the beta flag? There's a known problem with some recent CPU projects that sounds a lot like your description. If so, remove that setting.

[Volnaiskra]

No, I did run beta and/or adv earlier in the year, but have been on default for months now.

Hmm. I didn't know that GPUs can take up to a minute to stop. But I'm pretty sure that I've waited well over 5 minutes in the past. There may even have been times where it took me half an hour or so to realise that it was still folding, even though it was on "idle".

[Volnaiskra]

ok, I can now confirm that this problem doesn't seem related to my PC. My secondary PC gets the same problem: 'unsent' WUs piled up until there were about 6 of them, and now FAHcontrol is greyed out, and "idle" mode doesn't work. I had never seen this on the 2nd PC until now. It has a much weaker GPU, so maybe it's because it took longer to accumulate those unsent WUs.

All the evidence seems to point to my router, since that is the only shared piece of hardware/software between the two computers. Also, I bought this router in Dec 2013, which is roughly when these problems started happening.

Can someone who knows about routers please talk me through various settings I should look out for? As I've mentioned earlier, the router has a firewall feature but it's not being used. It also has the latest firmware, and my connection works perfectly in every other scenario I throw at it. I can take screenshots of the various pages of the router setup firmware and post them if someone can talk me through the details to look out for.