You are not logged in.
Is this still active? I joined in three days ago
Offline
For chrome/chromium users, there is a also app to run the software directly within your browser
Offline
OK, Buddlespit is back folding@home... After a year away, I'm still ranked in the top 30.
For chrome/chromium users, there is a also app to run the software directly within your browser
I couldn't get it to use/see my gpu.
Last edited by Buddlespit (2016-07-02 14:54:25)
Offline
I couldn't get it to use/see my gpu.
I'm guessing it is only CPU for the chrome app.
Offline
Is there a way to limit the CPU usage on a non-GUI system?
i have foldingathome-noroot AUR installed.
Last edited by cra1g321 (2016-08-06 00:03:21)
Offline
Is there a way to limit the CPU usage on a non-GUI system?
i have foldingathome-noroot AUR installed.
Define 'CPU usage'. You can limit the number of threads, how aggressive fah is as a background task or whether it runs only when on idle or not.
Offline
Define 'CPU usage'. You can limit the number of threads, how aggressive fah is as a background task or whether it runs only when on idle or not.
Like configure it to only use a certain percentage of the CPU power, rather than hitting 100% CPU usage constantly.
Offline
Buddlespit wrote:Define 'CPU usage'. You can limit the number of threads, how aggressive fah is as a background task or whether it runs only when on idle or not.
Like configure it to only use a certain percentage of the CPU power, rather than hitting 100% CPU usage constantly.
No. You can limit the number of threads, how aggressive fah is as a background task or whether it runs only when on idle or not.
Offline
No. You can limit the number of threads, how aggressive fah is as a background task or whether it runs only when on idle or not.
I did a bit of google searching but couldnt find anything on how to do this for a non-GUI system, if you could provide the answer or some links, it would be appreciated.
Offline
your /opt/fah/config.xml
<!-- Folding Slots -->
<!-- Use all the CPUs
Watch out for high load -->
<slot id='0' type='CPU'/>
<cpu v=2/>
<slot id='1' type='GPU'/>
cpu v=number of threads to use. I think you have to use an even number because fah will use the actual cores. I tried '3' on my 6 thread, 3 core AMD and it used 4 (two cores). 2 and 4 seem to work correctly (one and two cores).
<!-- Slot Control
Options: light, medium or full
Watch out for high load -->
<power v='full'/>
Self explanatory. Use either slot control or cpu v=. I don't think you can use both at the same time, but I've never tried it either.
Offline
Hi all,
I am wondering if there is a way to recover my passkey as I want to begin folding again but do not have it!
Thanks in advance,
Diesel1.
*EDIT* Found it!
http://fah-web.stanford.edu/cgi-bin/getpasskey.py
Last edited by diesel1 (2016-11-28 19:19:39)
Registered GNU/Linux user #140607.
Offline
Hi all,
I have noticed that my gpu is hardly ever running f@h work units. The web client seems to be constantly suggesting there will be a job to do sometime in the future but the gpu never seems to show an active job.
Have any of you seen this behaviour and do you have any suggestions as to what to do?
Diesel1.
*EDIT* - I have noticed that when I cycle the power settings on the web control page the client attempts to collect a new gpu work unit.
Last edited by diesel1 (2016-12-20 15:16:07)
Registered GNU/Linux user #140607.
Offline
Finally made a forum account, so of course I had to stop by and greet my fellow folders in team 45032. Also pretty surprising to see us so close to the Ubuntu team!
Hi all,
I have noticed that my gpu is hardly ever running f@h work units. The web client seems to be constantly suggesting there will be a job to do sometime in the future but the gpu never seems to show an active job.
Have any of you seen this behaviour and do you have any suggestions as to what to do?
Diesel1.
*EDIT* - I have noticed that when I cycle the power settings on the web control page the client attempts to collect a new gpu work unit.
Which GPU and drivers are you using? You could also try installing FAHControl and taking a look at the "Log" tab (this is probably also a text file somewhere on the system, but I find FAHControl quite handy), perhaps with the "Warnings & Errors" checkbox ticked.
Offline
Hello nassi!
Thanks for the reply. I was beginning to think I was alone on here.
I had forgotten about FAHControl but I will try it right now.
Diesel1.
Registered GNU/Linux user #140607.
Offline
Hi all,
This is what the log shows...
14:34:26:WU00:FS01:Assigned to work server 171.64.65.92
14:34:26:WU00:FS01:Requesting new work unit for slot 01: READY gpu:0:GP106 [GeForce GTX 1060 6GB] from 171.64.65.92
14:34:26:WU00:FS01:Connecting to 171.64.65.92:8080
14:34:27:WU00:FS01:Downloading 3.05MiB
14:34:33:WU00:FS01:Download 57.29%
14:34:35:WU00:FS01:Download complete
14:34:35:WU00:FS01:Received Unit: id:00 state:DOWNLOAD error:NO_ERROR project:9196 run:1 clone:58 gen:209 core:0x21 unit:0x0000013cab40415c57cb3f3392b642bd
14:34:35:WU00:FS01:Starting
14:34:35:WU00:FS01:Running FahCore: /opt/fah/FAHCoreWrapper /opt/fah/cores/fahwebx.stanford.edu/cores/Linux/AMD64/NVIDIA/Fermi/Core_21.fah/FahCore_21 -dir 00 -suffix 01 -version 704 -lifeline 560 -checkpoint 15 -gpu 0 -gpu-vendor nvidia
14:34:35:WU00:FS01:Started FahCore on PID 19721
14:34:35:WU00:FS01:Core PID:19725
14:34:35:WU00:FS01:FahCore 0x21 started
14:34:36:WU00:FS01:0x21:*********************** Log Started 2017-01-29T14:34:35Z ***********************
14:34:36:WU00:FS01:0x21:Project: 9196 (Run 1, Clone 58, Gen 209)
14:34:36:WU00:FS01:0x21:Unit: 0x0000013cab40415c57cb3f3392b642bd
14:34:36:WU00:FS01:0x21:CPU: 0x00000000000000000000000000000000
14:34:36:WU00:FS01:0x21:Machine: 1
14:34:36:WU00:FS01:0x21:Reading tar file core.xml
14:34:36:WU00:FS01:0x21:Reading tar file system.xml
14:34:36:WU00:FS01:0x21:Reading tar file integrator.xml
14:34:36:WU00:FS01:0x21:Reading tar file state.xml
14:34:36:WU00:FS01:0x21:Digital signatures verified
14:34:36:WU00:FS01:0x21:Folding@home GPU Core21 Folding@home Core
14:34:36:WU00:FS01:0x21:Version 0.0.17
14:34:44:WU00:FS01:0x21:ERROR:exception: Error downloading array interactionCount: clEnqueueReadBuffer (-5)
14:34:44:WU00:FS01:0x21:Saving result file logfile_01.txt
14:34:44:WU00:FS01:0x21:Saving result file log.txt
14:34:44:WU00:FS01:0x21:Folding@home Core Shutdown: BAD_WORK_UNIT
14:34:50:WARNING:WU00:FS01:FahCore returned: BAD_WORK_UNIT (114 = 0x72)
14:34:50:WU00:FS01:Sending unit results: id:00 state:SEND error:FAULTY project:9196 run:1 clone:58 gen:209 core:0x21 unit:0x0000013cab40415c57cb3f3392b642bd
14:34:50:WU00:FS01:Uploading 2.49KiB to 171.64.65.92
14:34:50:WU00:FS01:Connecting to 171.64.65.92:8080
14:34:50:WU02:FS01:Connecting to 171.67.108.45:80
14:34:50:WU00:FS01:Upload complete
14:34:50:WU00:FS01:Server responded WORK_ACK (400)
14:34:50:WU00:FS01:Cleaning up
14:34:51:WU02:FS01:Assigned to work server 140.163.4.244
I am not sure what the error means.
Diesel1.
Registered GNU/Linux user #140607.
Offline
I was getting the same error (BAD_WORK_UNIT (114 = 0x72)) on a AMD card a while back. It indicates a driver problem. Using the "clEnqueueReadBuffer (-5)" line in your log for googling reveals this to indeed be true: https://foldingforum.org/viewtopic.php?f=96&t=29384.
The current proprietary driver version in the repositories seems to be 375.26-6, so if you are using the latest proprietary driver then that is most likely the problem. A temporary driver downgrade might work (as suggested here https://foldingathome.stanford.edu/2017 … a-drivers/), but I'm not sure if it will break some dependencies.
Last edited by nassi (2017-01-29 15:50:48)
Offline
Thanks nassi,
I did read the same thing and I have noticed that all of the 0x18 units seem to work but the 0x21 units fail.
Diesel1.
Registered GNU/Linux user #140607.
Offline
Since I last updated the system F@H is running consistently.
Diesel1.
Registered GNU/Linux user #140607.
Offline
The Arch Linux Folding@Home team just overtook TeamUbuntu and at a glance this makes us the leader among the Linux teams of F@H! Whoop!
Offline
Good, we were up at 76 once upon a time.
Registered GNU/Linux user #140607.
Offline
Good, we were up at 76 once upon a time.
If only we had a few more 50K/day users...
Looking at the current list, I find it funny that we have more than half the worker count of Ubuntu with a slightly higher score. There might not be as many of us, but our computers are better. I only started folding for us a few days ago.
Not sure if I should post this here or in the AUR section of the forum...but anyways,
I've setup some PKGBUILD's for the latest F@H beta client based upon the current "foldingathome", "fahcontrol", & "fahviewer" packages in the AUR. These use version 7.4.16 instead of 7.4.4 from the current packages. Unfortunately, they don't install side-by-side one another because both FAHControl versions install the same files under /usr...some awk and sed magic might be able to overcome that, but I haven't looked into it yet. If one doesn't use FAHControl, both stable and beta can be installed along side one another, just remove the conflicts and replaces lines from my PKGBUILDs below. Beta is only available for X86_64.
https://github.com/skeevy420/foldingathome-beta-bin
https://github.com/skeevy420/fahcontrol-beta-bin
https://github.com/skeevy420/fahviewer-beta-bin
If these work for others I'm willing to submit 'em to the AUR and maintain them. I just don't want to submit a package that others might not want that's only been tested by myself for 24 hours.
Also going to add, because this might help others, regardless of the F@H version, the package "opencl-amd" with AMDGPU-PRO OpenCL from the AUR is necessary for folding to work with the AMDGPU mesa driver...at least with CIK enabled kernels using a CIK device like my R7 260x. I've since set it up to only use that implementation of OpenCL -- for me, it was set the opencl-index of my GPU to 0.
Offline
Hi skeevy420,
Welcome to the F@H Arch Linux Team.
Do the beta packages produce more PPD?
Do your pkg builds work with Nvidia cards or are they tuned for AMD?
Kind regards,
Diesel1.
Registered GNU/Linux user #140607.
Offline
Not sure on the PPD. I haven't used either long enough to make an accurate comparison, but I think my GPU is getting an extra 2000 points per 15 hours. Currently, I only have 130K points (all to Arch) over 2 days of on and off use with the stable client. I'll know more in another 36 hours of use with the beta client.
I only have AMD GPU's so I can't say for certain that NVidia will work, but I don't see any reason why NVidia GPUs won't.
They're not tuned for anything, just pointing towards newer versions of the same packages the current stable clients are using. The only differences are I removed x86 support (not available), added "opencl-amd" as an optional dependency (actually required for me), and set them to conflict with stable packages because FAHControl stable and beta won't install side-by-side.
Other than that, they're the same and the same instructions apply. "makepkg -sri", cd to "/opt/fah-beta", do the "./FAHClient --configure" step, and enable/start the foldingathomebeta.service for systemd.
I did notice a massive 50% PPD increase for my CPU, but I think that's because I set my threads from 15 to 16 and my "client-type" to "bigadv" than the change in clients. I have a dual X5687 system with 16 threads and I figured I might as well use the 16+ thread mode. I didn't do that until this morning. From what I've read, "bigadv" projects give out more points and I went from ~4K with 2 hour projects to ~16K with 4 hour projects.
Offline
@skeevy420
I tried your beta drivers and it upset my system no end (bad skeevy420!!)! I also fell foul of the CA certificates issue and ended up reinstalling. Since then I am unable to fold with the GPU as it is failing on the 0x21 units. I cannot remember how I made the client only run 0x18 units!
As it stands, it would seem I need a major downgrade of the nVidia driver or a miracle (I am a devout ((sic)) atheist!) so the downgrade seems most likely, just how to achieve it and is it worth it?
Is there a config option to only download 0x18 work units?
Diesel1.
Registered GNU/Linux user #140607.
Offline
@diesel1
I don't think you can choose what WUs you'll receive, as it would make people only accept the WUs with the best PPD returns. As to the current state of the nVidia drivers, I couldn't get a GTX 680 to fold (under Windows 10, was selling the machine so didn't bother trying it out on Arch) using the latest drivers (had to use 373.XX or something equally old). Summer is right around the corner though, I probably wouldn't bother messing with it (also partly because I want to hold my lead over you in the rankings ).
@skeevy420
I was under the impression that bigadv was discontinued quite a while back. Maybe what you were seeing was a combination of the boost provided by going from 15 to 16 threads in combination with a better-than-average WU? See for example this: https://foldingforum.org/viewtopic.php?f=96&t=29824.
Haven't had my main rigs folding for a while now (290X and 390X do provide nice PPD, but also quite a bit of heat and power draw). Might just sell them and get something more power efficient.
Last edited by nassi (2017-04-07 23:10:55)
Offline