Message boards :
Questions and problems :
How to use GPU?
Message board moderation
Previous · 1 · 2
Author | Message |
---|---|
Send message Joined: 15 Apr 15 Posts: 50 |
Thanks, got the 128 core box back to running 50 threads. Still not where I need it to be, but much better than before. Seems to be a limitation of the boinc_client program. Checked command line options, and there is one which seems to allow multiple clients, however, when I tried to run 2 instances of boinc_client, it just complained that one was already running, even with that option. |
Send message Joined: 23 Apr 07 Posts: 1112 |
Seems to be a limitation of the boinc_client program. Checked command line options, and there is one which seems to allow multiple clients, however, when I tried to run 2 instances of boinc_client, it just complained that one was already running, even with that option. Did you make another directory location, and install Boinc in it, then run that instance with the allow allow multiple client? each instance will need to a have a different host id, etc, so you can't run a client multiple times from the same directory. Claggy |
Send message Joined: 29 Aug 05 Posts: 15483 |
Seems to be a limitation of the boinc_client program. Because it couldn't possibly be a limitation set elsewhere, or yet again a bad preference setting. Without messages what this BOINC does at such a moment, there's not much I can see about it either, but I can tell you that the BOINC client has no limitation on CPU cores it can detect and use. As long as the hostID is seen as reliable, your work + additional work request is set to something high enough, and the project allows that much work to be sent to your client, it can stock up work and run it simultaneously. So perhaps it's a project limitation, or it's something else. Impossible to guess without you giving some message lines of a scheduler contact with the project, preferably with the <sched_op_debug> flag set on. You may also want to read When requesting help on these forums for a clue of the bare minimum we like to get when guessing at answers, without constantly having to play 20 questions. As for multiple_clients, and as Claggy asked, you did set them up with different data directories and when paired against a BOINC Manager, different port numbers? Something like: CD /D c:\Boinc1 "C:\Program Files\BOINC\boinc.exe" --detach --allow_multiple_clients --gui_rpc_port 31417 --dir c:\BOINC1 "C:\Program Files\BOINC\boincmgr.exe" /m /n localhost /g 31417 /p xxxxxxxxxx CD /D c:\Boinc2 "C:\Program Files\BOINC\boinc.exe" --detach --allow_multiple_clients --gui_rpc_port 31419 --dir c:\BOINC2 "C:\Program Files\BOINC\boincmgr.exe" /m /n localhost /g 31419 /p xxxxxxxxxx (Index: FAQ on BOINC command line switches, FAQ on BOINC Manager command line switches) |
Send message Joined: 15 Apr 15 Posts: 50 |
No, I wasn't aware that separate directories were needed. Ill try that and see if I can get it going at 100 threads instead of 50. As for the host side, I know the host will support all 128 threads, as it was running a database engine prior to being replaced with a newer/larger box. So I do believe it's something on the application side, but since that includes all the application settings, there could be a limitation in those settings I haven't discovered yet. Ill try and read more about the various settings and what they control. |
Send message Joined: 15 Apr 15 Posts: 50 |
Tried multiple directories, but no luck. The second instance doesn't run. I have also not been successfull in locating any option settings which limit the boinc instance to 50, so im thinking it may be an inherent limitation. Will have to ask on Boinc support forum to get more details on that issue. 21-Apr-2015 06:43:40 [---] Starting BOINC client version 7.2.33 for x86_64-pc-linux-gnu 21-Apr-2015 06:43:40 [---] log flags: file_xfer, sched_ops, task 21-Apr-2015 06:43:40 [---] Libraries: libcurl/7.19.7 NSS/3.16.2.3 Basic ECC zlib/1.2.3 libidn/1.18 libssh2/1.4.2 21-Apr-2015 06:43:40 [---] Data directory: /home/rwatkins/.hp6b 21-Apr-2015 06:43:40 [---] No usable GPUs found 21-Apr-2015 06:43:40 [---] Host name: hp6.dssolutions.com 21-Apr-2015 06:43:40 [---] Processor: 128 GenuineIntel Intel(R) Xeon(R) CPU X6550 @ 2.00GHz [Family 6 Model 46 Stepping 6] 21-Apr-2015 06:43:40 [---] Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx rdtscp lm constant_tsc arch_perfmon pebs bts rep_good xtopology nonstop_tsc aperfmperf pni dtes64 monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm dca sse4_1 sse4_2 x2apic popcnt lahf_lm ida epb dts tpr_shadow vnmi flexpriority ept vpid 21-Apr-2015 06:43:40 [---] OS: Linux: 2.6.32-431.el6.x86_64 21-Apr-2015 06:43:40 [---] Memory: 504.71 GB physical, 4.00 GB virtual 21-Apr-2015 06:43:40 [---] Disk: 188.99 GB total, 69.35 GB free 21-Apr-2015 06:43:40 [---] Local time is UTC -7 hours 21-Apr-2015 06:43:40 [---] Config: don't use coprocessors 21-Apr-2015 06:43:40 [---] Config: report completed tasks immediately 21-Apr-2015 06:43:40 [---] Config: GUI RPC allowed from any host 21-Apr-2015 06:43:40 [---] A new version of BOINC is available. (7.2.42) <a href=http://boinc.berkeley.edu/download.php>Download</a> 21-Apr-2015 06:43:40 [---] Already attached to http://milkyway.cs.rpi.edu/milkyway/ 21-Apr-2015 06:43:40 [Milkyway@Home] URL http://milkyway.cs.rpi.edu/milkyway/; Computer ID 617118; resource share 100 21-Apr-2015 06:43:40 [Milkyway@Home] General prefs: from Milkyway@Home (last modified 21-Apr-2015 04:36:39) 21-Apr-2015 06:43:40 [Milkyway@Home] Host location: none 21-Apr-2015 06:43:40 [Milkyway@Home] General prefs: using your defaults 21-Apr-2015 06:43:40 [---] Preferences: 21-Apr-2015 06:43:40 [---] max memory usage when active: 258410.03MB 21-Apr-2015 06:43:40 [---] max memory usage when idle: 465138.06MB 21-Apr-2015 06:43:40 [---] max disk usage: 68.36GB 21-Apr-2015 06:43:40 [---] suspend work if non-BOINC CPU load exceeds 25% 21-Apr-2015 06:43:40 [---] (to change preferences, visit a project web site or select Preferences in the Manager) 21-Apr-2015 06:43:40 [---] GUI RPC bind to port 31416 failed: 98 21-Apr-2015 06:44:10 gstate.init() failed Error Code: -180 Anyone know what the "failed" message is about? Does this have somthing to do with trying to run 2 boinc_client processes? |
Send message Joined: 15 Apr 15 Posts: 50 |
Ive been a bad boy and let this thread drift off topic. The initial issue regarding GPU's is resolved. My 7850's are running, although VERY slowly compared to CPUs. Ill take the cpu threading issue to a new topic. |
Copyright © 2024 University of California.
Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License,
Version 1.2 or any later version published by the Free Software Foundation.