Posts by Werkstatt
|
1)
Message boards :
Number crunching :
Errors
(Message 66649)
Posted 18 Sep 2017 by  Werkstatt
Post: Hey Everyone, Sorry, I still get bad wu's. |
|
2)
Message boards :
Number crunching :
Errors
(Message 66638)
Posted 17 Sep 2017 by Werkstatt
Post: I've sent a note to the admins. |
|
3)
Message boards :
Number crunching :
Errors
(Message 66616)
Posted 16 Sep 2017 by Werkstatt
Post: Add me too http://milkyway.cs.rpi.edu/milkyway/results.php?userid=2765&offset=0&show_names=0&state=6&appid= Some of these wu's fail on other systems too. |
|
4)
Message boards :
Number crunching :
Updated GPU Requirements (Currently not supporting GPU tasks)
(Message 66012)
Posted 13 Dec 2016 by Werkstatt
Post: I tried this on my GTX-750ti in Ubuntu 14.04 and get the following. I am running 334.21 which as far as I know is the only driver for this card in Ubuntu at this time. The card runs Einstein, DiRT and Collatz fine. But it does not run Folding@Home currently. Is this just a matter of waiting for the developers to provide work units that support Maxwell, or is there something I need to do? I can run MW on my GTX 750ti, using win7. Your message looks like you have no open-CL package installed. I would start there when looking for a solution. |
|
5)
Message boards :
News :
Scheduled Maintenance Concluded
(Message 65951)
Posted 23 Nov 2016 by Werkstatt
Post:
That's easy. With an app_config.xml you can set the cpu and gpu usage. I've set Einstein and Gpugrid to 0.34 gpu's, which means, it can run two wu's. And I've set MW to use 0.25 gpu's, so the BM loads another app to fill the gpu usage to the best value near 100%. The actual mix of apps ist then controlled by the ressource share settings. When no MW wu's are available, it will load Einstein for AMD and Gpugrid for nVidia, since ressource share is set to 0. If you use the same share for all projects, the mix changes from time to time. With the app_config as described you have always at least one MW wu per gpu running. Edit: sorry its not app_info, its app_config.xml |
|
6)
Message boards :
News :
Scheduled Maintenance Concluded
(Message 65949)
Posted 23 Nov 2016 by Werkstatt
Post: What I find an absolute nightmare is trying to get AMD drivers to load when I also have NVidia GPU running:-( Dunno what the secret is but its eluded me.. Absolutely no problem, look at this pc: http://milkyway.cs.rpi.edu/milkyway/show_host_detail.php?hostid=581360 It's not a problem to mix cards and it's not a problem to mix apps. I can run GPUGRID and MW at the same time on the nVidea card, Einstein and MW on AMD and Einstein on Intel. https://dl.dropboxusercontent.com/u/50246791/crunching%202%20on%20mainsys%204.PNG Alexander |
|
7)
Message boards :
Number crunching :
Android/ARM WUs all failing with computation errors
(Message 65846)
Posted 15 Nov 2016 by Werkstatt
Post: I'm not really sure how practical it is to use these mobile devices anyway. Other than not being very powerful compared to full higher end PCs, they're specifically made and optimized to be mostly idle most of the time and aren't suited to being up and crunching in the background for extended periods of time. No. 14 months ago I added an Odroid to the project, an overclocked to 4x2 GHz Arm V7, crunching 24/7 MW only. It took less than 5 weeks to see this device in the top 10,000 by overall credit! I have my very first tablet in the kitchen as a radio, it did also crunch MW until the app did no longer work, also 24/7. I have a 8" tablet, that was fallen down; 40% of the touch area is no longer working as a touch, but it works 24/7 as a crunching device, currently only for the skynet pogs. It has an average credit of 1,405 an an overall of 467,900. Much better than throwing it away! I love android wu's! |
|
8)
Message boards :
News :
Scheduled Maintenance Concluded
(Message 65804)
Posted 14 Nov 2016 by Werkstatt
Post: Great work, Jake, works fine on three pc's now. |
|
9)
Message boards :
Number crunching :
Tracking in Windows 7
(Message 65774)
Posted 14 Nov 2016 by Werkstatt
Post:
I tried it twice. Was unable (also with help from a forum) to install AMD openCL. It worked easy with nVidia. I followed a three month discussion @ Einstein. I never had a crunching AMD Linux version. No, thanks. Alexander |
|
10)
Message boards :
Number crunching :
Tracking in Windows 7
(Message 65751)
Posted 13 Nov 2016 by Werkstatt
Post: Ich sehe gerade, dass Sie wohl deutschsprachig sind. Klicken Sie auf "Start", geben Sie unten "Dienste" ein und betätigen Sie die Eingabetaste ;) Besten Dank! Thanks a lot! |
|
11)
Message boards :
Number crunching :
Tracking in Windows 7
(Message 65735)
Posted 13 Nov 2016 by Werkstatt
Post: Where can i find that? |
|
12)
Message boards :
News :
Scheduled Maintenance Concluded
(Message 65684)
Posted 11 Nov 2016 by Werkstatt
Post: Well, yes, they start crunching, but at ~70% they restart with 0%. In 45 minutes not a single gpu wu finished. |
|
13)
Message boards :
Number crunching :
Server error: feeder not running
(Message 65604)
Posted 6 Nov 2016 by Werkstatt
Post: The problems have shifted a little bit. At the moment the work generation is the bottleneck. |
|
14)
Message boards :
Number crunching :
Android/ARM WUs all failing with computation errors
(Message 65601)
Posted 6 Nov 2016 by Werkstatt
Post: Unfortunately they do not work anymore. The project was updated some months ago and the old arm wu's all fail. Believe me, i am sad too. |
|
15)
Message boards :
News :
Updated Server Daemons and Libraries
(Message 65588)
Posted 3 Nov 2016 by Werkstatt
Post: Hi Sebastian* Yes, I know this problem. I have this on my PC with the HD58xx card, when running Einstein and MW mixed. Sometimes a message like 'AMD driver reset after driver error' is displayed. Other systems accept these settings, but with older drivers. The 'HD58xx-System' is a 'repaired win7' after many crashes and an abstruse mix of programs and browser add on, cleaned but never prooved to be as 'out of factory'. This is why I never payed too much attention on it. But works with 4 MW simultanously. Check my systems for driver versions, currently all my systems work fine with MW. Alexander |
|
16)
Message boards :
Number crunching :
How can I increase my workbuffer
(Message 65582)
Posted 2 Nov 2016 by Werkstatt
Post: Thank you, a clear answer! Good luck for your work! |
|
17)
Message boards :
Number crunching :
HD48xx not getting work
(Message 65580)
Posted 2 Nov 2016 by Werkstatt
Post: Are these problems solved? If yes I will rebuild one PC to run 2 of these gpu's. |
|
18)
Message boards :
Number crunching :
How can I increase my workbuffer
(Message 65579)
Posted 2 Nov 2016 by Werkstatt
Post: I'm crunching gpu wu's only. I've set my computers to store wu's for 5 days and an additional buffer also of 5 days. Hours later I had wu's for about one hour. I want a buffer for 4 to 5 hours to keep my gpu's active during the many short database outages. Any ideas beside crunching other projects? Alexander |
|
19)
Message boards :
Number crunching :
HD6970 & Newer versions of WUs
(Message 65450)
Posted 15 Oct 2016 by Werkstatt
Post: OK, thanks! So it looks like that's not an issue anymore. No, it's 15.09 |
|
20)
Message boards :
Number crunching :
HD6970 & Newer versions of WUs
(Message 65444)
Posted 14 Oct 2016 by Werkstatt
Post: Check this pc: http://milkyway.cs.rpi.edu/milkyway/results.php?hostid=486169 Edit: running 4 wu's at a time |
Next 20
©2024 Astroinformatics Group