Message boards :
Number crunching :
NOmore CPU crunshing
Message board moderation
Author | Message |
---|---|
The_Matrix Send message Joined: 17 Nov 03 Posts: 414 Credit: 5,827,850 RAC: 0 |
guess what, my pc dont crunsh CPU tasks anymore. the client says OpenCL CPU detected, thats all i know. Using the newest dev client. HELP ! |
Thomas Send message Joined: 9 Dec 11 Posts: 1499 Credit: 1,345,576 RAC: 0 |
Check your SETI@home preferences on your account (USE CPU = yes) Maybe it's necessary to downgrade to 7.2.42 |
The_Matrix Send message Joined: 17 Nov 03 Posts: 414 Credit: 5,827,850 RAC: 0 |
ok, i just trying. the app switch was at 100.000 minutes, now back at 60. hope that will help, but when not i will downgrade. THX So i cant wait, i downgraded the client, everythings fine. THX for the help. |
The_Matrix Send message Joined: 17 Nov 03 Posts: 414 Credit: 5,827,850 RAC: 0 |
i am back again, same issue. On the older client software. I got enough wus to crunsh , but no cpu crunshing at all. What should/can i do ? -.- no solution ? |
The_Matrix Send message Joined: 17 Nov 03 Posts: 414 Credit: 5,827,850 RAC: 0 |
Still no automated cpu crunshing... -.- i will end computations , and then REMOVE the project from the client-manager. will report in a few hours... |
Zalster Send message Joined: 27 May 99 Posts: 5517 Credit: 528,817,460 RAC: 242 |
Matrix, what is your set up on that computer? How many work units are you crunching on the GPUs and are you using an app_info.xml or app_config.xml? Post the first 30 lines of the event log so we can look and see what it says Edit I only ask as GPU crunching is faster than the CPU. I have 2 machines where I only crunch on the GPUs and use the CPU to support them. They produce higher RAC rather than doing a combined CPU/GPU set up. |
The_Matrix Send message Joined: 17 Nov 03 Posts: 414 Credit: 5,827,850 RAC: 0 |
01.08.2014 16:32:21 | | Starting BOINC client version 7.2.42 for windows_x86_64 01.08.2014 16:32:21 | | log flags: file_xfer, sched_ops, task 01.08.2014 16:32:21 | | Libraries: libcurl/7.25.0 OpenSSL/1.0.1 zlib/1.2.6 01.08.2014 16:32:21 | | Data directory: I:\BOINC64\data 01.08.2014 16:32:21 | | CUDA: NVIDIA GPU 0: GeForce GTX 670 (driver version 337.50, CUDA version 6.0, compute capability 3.0, 2048MB, 1876MB available, 3372 GFLOPS peak) 01.08.2014 16:32:21 | | CUDA: NVIDIA GPU 1: GeForce GT 640 (driver version 337.50, CUDA version 6.0, compute capability 3.0, 1024MB, 959MB available, 692 GFLOPS peak) 01.08.2014 16:32:21 | | OpenCL: NVIDIA GPU 0: GeForce GTX 670 (driver version 337.50, device version OpenCL 1.1 CUDA, 2048MB, 1876MB available, 3372 GFLOPS peak) 01.08.2014 16:32:21 | | OpenCL: NVIDIA GPU 1: GeForce GT 640 (driver version 337.50, device version OpenCL 1.1 CUDA, 1024MB, 959MB available, 692 GFLOPS peak) 01.08.2014 16:32:21 | | OpenCL CPU: Intel(R) Core(TM) i5-3550 CPU @ 3.30GHz (OpenCL driver vendor: Intel(R) Corporation, driver version 1.1, device version OpenCL 1.1 (Build 31360.31441)) 01.08.2014 16:32:21 | | Processor: 4 GenuineIntel Intel(R) Core(TM) i5-3550 CPU @ 3.30GHz [Family 6 Model 58 Stepping 9] 01.08.2014 16:32:21 | | 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 htt tm pni ssse3 cx16 sse4_1 sse4_2 popcnt aes syscall lm vmx smx tm2 pbe 01.08.2014 16:32:21 | | OS: Microsoft Windows 7: Professional x64 Edition, Service Pack 1, (06.01.7601.00) 01.08.2014 16:32:21 | | Memory: 11.95 GB physical, 23.89 GB virtual 01.08.2014 16:32:21 | | Disk: 465.76 GB total, 159.42 GB free 01.08.2014 16:32:21 | | Local time is UTC +2 hours 01.08.2014 16:32:21 | Poem@Home | Found app_config.xml 01.08.2014 16:32:21 | Einstein@Home | Found app_config.xml 01.08.2014 16:32:21 | Milkyway@Home | Found app_config.xml 01.08.2014 16:32:21 | SETI@home | Found app_config.xml 01.08.2014 16:32:21 | | Config: simulate 4 CPUs 01.08.2014 16:32:21 | | Config: report completed tasks immediately 01.08.2014 16:32:21 | | Config: use all coprocessors 01.08.2014 16:32:21 | | Config: fetch on update The only i noticed was , that the cpu usage have rised per gpu wu, this must be the reason for blocking cpu wus !? What do u mean ? |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
Note that your OpenCL detection for CPU is irrelevant for this project (and most other BOINC projects, I suspect). This project will only supply native-mode CPU applications, if your configurations and preferences are set up to ask for those. It does appear that your CPU contains an integrated Intel® HD Graphics 2500 GPU, but you would need to install a GPU OpenCL v1.2 driver (and jump through sundry other hoops, like attaching a monitor and/or extending your desktop) to make use of that. This project does have iGPU apps available, once the HD 2500 is enabled and recognised. |
The_Matrix Send message Joined: 17 Nov 03 Posts: 414 Credit: 5,827,850 RAC: 0 |
ok, restarting the project, worked so far. dont know 100% exactly what it was. The HD Grafphics 2500 GPU isnt online , not used by me. Because i cant play games well with that chip. 4 GPU , and 4 cpu tasks are running now. I hope that was it. |
Zalster Send message Joined: 27 May 99 Posts: 5517 Credit: 528,817,460 RAC: 242 |
Ok, since the last time you post till now you computer contacted seti and I can see it now has some CPU work units. So, let it run and complete those CPU work units and go from there. |
juan BFP Send message Joined: 16 Mar 07 Posts: 9786 Credit: 572,710,851 RAC: 3,799 |
4 GPU , and 4 cpu tasks are running now. I hope that was it. May i sugest you to check the number of CPU tasks, 4 GPU + 4 CPU tasks on a I5 who only have 4 cores is maybe too much, specialy if you run any AP tasks. I have a 2x670 host powered by a I5 too, but my I5 is older & slower than yours. In my case the optimal numbers is when it run 2 or less CPU WU and 4 GPU WU. My host easely handle up to 4 AP GPU WU at a time, but in this case it can´t do any CPU work or the entire system slow down. On the other hand, when i crunch MB only CPU work is welcomed. So you need to adjust the config file to automatcaly adjust the number of CPU work depending of the type of WU you where crunching. Of course YMMV and test is required as usual. Just a tip. |
Mike Send message Joined: 17 Feb 01 Posts: 34257 Credit: 79,922,639 RAC: 80 |
I totally agree with Juan here. With each crime and every kindness we birth our future. |
The_Matrix Send message Joined: 17 Nov 03 Posts: 414 Credit: 5,827,850 RAC: 0 |
No furver issues. Still running 4 cpu and 4 gpu tasks, no browsing performance failures. : D thats all i need while crunshing. |
juan BFP Send message Joined: 16 Mar 07 Posts: 9786 Credit: 572,710,851 RAC: 3,799 |
Is not about failures, is about optimization. Your GPUs are a lot faster than the CPU for crunching, so to optimize our cruncher you need to first keep your GPU´s well feeded, and to do that, specialy on less cores/powerfull CPU like the I5 we have, you need no keep some cores free to do that job. Let me explain ussing my host as example, This host: http://setiathome.berkeley.edu/show_host_detail.php?hostid=7347263 has 2x670 and a I5-2310 CPU @ 2.90GHz no OC. It´s running now 3 MB WU at a time on each GPU, at this moment it´s running a total of 6 MB GPU , i could power up to 2 CPU WU and the GPU usage will remain at close to 100%, if i start 4 for example the GPu usage drops to about 80%, why? simply the GPU loose time waiting for the CPU to feed it. Then less GPU usage = less WU (CPU+GPU) crunched per hour. Now let´s change to AP, when it´s runs 3 AP GPU WU at a time on each GPU, (for a total of 6 on both GPUs) i can´t start a single CPU WU or the GPU usage drops to about 60% or even less. So 4 or more AP GPU AP means 0 CPU WU simple like that. Now you could ask, why i run 6 AP WU on the GPU´s if i only have 4 cores, i use a newer build (still on beta test) who uses less CPU than the stock ones, but that is for another time. So that is the unwrited rule, works with allmost all hosts but is more important to follow on slow CPU hosts who drives fast GPU´s, keep 1 CPU core for each AP GPU WU you where running so 4 AP WU in a 4 core I5 means No CPU work, or in the case of MB GPU keep at least 1 core free to feed the GPU´s (prefearely 2 on a 2xGPu hosts like ours), so with 4 GPU MB running use a max of 3 MB CPU WU (prefearely only 2). Of course YMMV and each host is unique that´s why i say test is needed, if you looking for better performance (more RAC inclusive). Some hosts can handle to feed the GPU´s with less cores, specialy the most powerful CPU´s but the roule works in the vast majority of hosts. BTW: Not forget, better GPU usage mean more heat dissipate, so you need to keep an eye on the GPU temperatures, and use a program to increse the fan speed, I use EVGA Precision (Free to DL from EVGA site and works on most NV no mather the brand) but there are others who do the same trick from others brands like MSI Afterburner, etc. |
Zalster Send message Joined: 27 May 99 Posts: 5517 Credit: 528,817,460 RAC: 242 |
There is a problem with the latest Version of Precision X. They pulled it from the website. If you go that route get an older version. I'm still running Version 4.2.1 Last time I checked the EVGA website the download links were disabled. |
The_Matrix Send message Joined: 17 Nov 03 Posts: 414 Credit: 5,827,850 RAC: 0 |
ok feedback. From itself , it has it normalysed. Automaticly it runs now at 3 cpu and 4 gpu tasks. |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.