Questions and Answers :
Unix/Linux :
Benchmarks started noticing only one processor.
Message board moderation
Author | Message |
---|---|
![]() Send message Joined: 10 Jun 99 Posts: 60 Credit: 1,301,105 RAC: 1 ![]() |
My machine runs Red Hat Enterprise Linux 3 with kernel 2.4.21-37.0.1.ELsmp. It has two hyperthreaded Xeon processors, so it acts as though it has 4 processors. Until about two days ago, it ran fine, but at one point it ran the benchmarks, detected only one processor, and uses only one processor: Mon Feb 20 09:55:44 2006||Suspending computation and network activity - running CPU benchmarks Mon Feb 20 09:55:44 2006|climateprediction.net|Pausing result sulphur_hral_100828525_0 (removed from memory) Mon Feb 20 09:55:46 2006||Running CPU benchmarks Mon Feb 20 09:55:46 2006||request_reschedule_cpus: process exited Mon Feb 20 09:56:45 2006||Benchmark results: Mon Feb 20 09:56:45 2006|| Number of CPUs: 1 Mon Feb 20 09:56:45 2006|| 826 double precision MIPS (Whetstone) per CPU Mon Feb 20 09:56:45 2006|| 1848 integer MIPS (Dhrystone) per CPU Mon Feb 20 09:56:45 2006||Finished CPU benchmarks Mon Feb 20 09:56:46 2006||Resuming computation and network activity I tried stopping and starting the BOINC client, and that did not help. I rebooted the system, and that did not help either. I reloaded everything in the /boinc directory for a 2-day old backup (made before I saw the problem), and it runs for a while, but when it runs the benchmarks, it discovers only one processor and uses only one. The computer is running all 4 processors on various tasks; i.e., the processor did not get hyperthreading turned off and switch to a uniprocessor kernel. BOINC Client is from this file: boinc_5.2.13_i686-pc-linux-gnu.sh What do I do now? ![]() |
Dotsch ![]() Send message Joined: 9 Jun 99 Posts: 2422 Credit: 919,393 RAC: 0 ![]() |
Strange. Can you please post the messages of the output from the startup of the boinc client (including the Processor/Memory and Disk output). If I have seen right in your computer stats, then will the boinc client detect the four CPUs. But strange, that the benchmark reports one. |
Temujin Send message Joined: 19 Oct 99 Posts: 292 Credit: 47,872,052 RAC: 0 ![]() |
strange one, only things I can suggest are 1, have you set your preferences to only use 1 processor? 2, have you tried running the benchmarks again? |
![]() Send message Joined: 10 Jun 99 Posts: 60 Credit: 1,301,105 RAC: 1 ![]() |
Strange. Can you please post the messages of the output from the startup of the boinc client (including the Processor/Memory and Disk output). I was going to post what you asked, but by the time I noticed these replies I had done an extreme restart: 1.) unmounted the /boinc file system 2.) mke2fs on top of it 3.) remounted /boinc file system 4.) copied in boinc_5.2.13_i686-pc-linux-gnu.sh and did what needed to be done With it so clean, I knew it could not be a problem on my machine. So I looked at, among other things, and found this: $ cat global_prefs.xml <global_preferences> <source_project>http://climateprediction.net/</source_project> <source_scheduler>http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi</source_scheduler> <mod_time>1140444216</mod_time> <run_if_user_active/> <idle_time_to_run>3</idle_time_to_run> <cpu_scheduling_period_minutes>60</cpu_scheduling_period_minutes> <work_buf_min_days>0.1</work_buf_min_days> <max_cpus>1</max_cpus> <---<<< <disk_interval>60</disk_interval> <disk_max_used_gb>100</disk_max_used_gb> <disk_max_used_pct>95</disk_max_used_pct> <disk_min_free_gb>0.1</disk_min_free_gb> <vm_max_used_pct>75</vm_max_used_pct> <max_bytes_sec_down>0</max_bytes_sec_down> <max_bytes_sec_up>0</max_bytes_sec_up> </global_preferences> Now the line with the arrow <---<<< is clearly wrong and would account for the problem. But how did it get that way since the global preferences were set, I believed, to 4 ever since I started running the BOINC client last year sometime. Well, after the restart, I only attached to the climateprediction one, instead of all four (thank goodness) that I usually run, and there is the problem. So I went to their site and saw that the 4 had transformed itself into a 1. Now either I fumble-fingered it when doing something else (but I am not up there much), or their machine had a hiccup. So I fixed it back to 4 up there. I will check the other three application sites before attaching to them again, but my guess is that they will be OK. This would account for it running erratically. When the other applications run first, I was using all 4 processors, but when clientprediction started running (or maybe it was when it sent a trickle in), it dropped down to one. These computers can sure be perplexing at times. ;-) ![]() |
©2025 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.