Hello, I am currently running the following:

4/15/2012 8:37:22 PM | | Starting BOINC client version 7.0.25 for windows_x86_64
4/15/2012 8:37:22 PM | | log flags: file_xfer, sched_ops, task
4/15/2012 8:37:22 PM | | Libraries: libcurl/7.21.6 OpenSSL/1.0.0d zlib/1.2.5
4/15/2012 8:37:22 PM | | Data directory: C:\ProgramData\BOINC
4/15/2012 8:37:22 PM | | Running under account Will
4/15/2012 8:37:22 PM | | Processor: 4 GenuineIntel Intel(R) Core(TM) i5-2500K CPU @ 3.30GHz [Family 6 Model 42 Stepping 7]
4/15/2012 8:37:22 PM | | Processor: 256.00 KB cache
4/15/2012 8:37:22 PM | | 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 syscall nx lm vmx tm2 popcnt aes pbe
4/15/2012 8:37:22 PM | | OS: Microsoft Windows 7: Ultimate x64 Edition, (06.01.7600.00)
4/15/2012 8:37:22 PM | | Memory: 7.51 GB physical, 15.02 GB virtual
4/15/2012 8:37:22 PM | | Disk: 465.66 GB total, 371.77 GB free
4/15/2012 8:37:22 PM | | Local time is UTC -6 hours
4/15/2012 8:37:22 PM | | NVIDIA GPU 0: GeForce GTX 260 (driver version 296.10, CUDA version 4.20, compute capability 1.3, 896MB, 757MB available, 907 GFLOPS peak)
4/15/2012 8:37:22 PM | | OpenCL: NVIDIA GPU 0: GeForce GTX 260 (driver version 296.10, device version OpenCL 1.0 CUDA, 896MB, 757MB available)
4/15/2012 8:37:22 PM | | Config: don't compute while steam.exe is running
4/15/2012 8:37:22 PM | Einstein@Home | URL http://einstein.phys.uwm.edu/; Computer ID 4816521; resource share 100
4/15/2012 8:37:22 PM | SETI@home | URL http://setiathome.berkeley.edu/; Computer ID 6462036; resource share 100
4/15/2012 8:37:22 PM | World Community Grid | URL http://www.worldcommunitygrid.org/; Computer ID 1919783; resource share 100
4/15/2012 8:37:22 PM | SETI@home | General prefs: from SETI@home (last modified 09-Apr-2012 23:20:26)
4/15/2012 8:37:22 PM | SETI@home | Computer location: home
4/15/2012 8:37:22 PM | | General prefs: using separate prefs for home
4/15/2012 8:37:22 PM | | Reading preferences override file
4/15/2012 8:37:22 PM | | Preferences:
4/15/2012 8:37:22 PM | | max memory usage when active: 3846.55MB
4/15/2012 8:37:22 PM | | max memory usage when idle: 6923.80MB
4/15/2012 8:37:22 PM | | max disk usage: 15.00GB
4/15/2012 8:37:22 PM | | don't use GPU while active
4/15/2012 8:37:22 PM | | suspend work if non-BOINC CPU load exceeds 45 %
4/15/2012 8:37:22 PM | | (to change preferences, visit the web site of an attached project, or select Preferences in the Manager)
4/15/2012 8:37:22 PM | | Not using a proxy

I have my preferences set to use 100% of Multi-CPU and 100% CPU Time when computing. I also have it set to switch between applications every 120 minutes.

I have noticed that World Community Grid seems to always be using all four processors and was to understand that each project would use 33.3% of the CPU time and processors when computing. Is this right or do I hav a setting wrong?

Is there a way to make sure each project uses one Processor at a time?

Thank you for your time,
Will

I have my preferences set to use 100% of Multi-CPU and 100% CPU Time when computing. I also have it set to switch between applications every 120 minutes.

I have noticed that World Community Grid seems to always be using all four processors and was to understand that each project would use 33.3% of the CPU time and processors when computing. Is this right or do I hav a setting wrong?

Is there a way to make sure each project uses one Processor at a time?

Thank you for your time,
Will

Are you wanting boinc to only run on 1 processor or have all 4 processors running but each with a different project?

If the first then set to 25% of multi cpu, if the second then AFAIK it's not possible to enforce - the boinc scheduler decides what and when it will run - so sometimes you will find it running only one project on all 4 processors other times it might be running two, three or four projects - it depends on what the scheduler decides it needs to run.

---

CIC1=CC=C(C2=N[C@@H](CC(OC(C)(C)C)=O)C3=NN=C(C)N3C4=C2C(C)=C(C)S4)C=C1

[quote]

Are you wanting boinc to only run on 1 processor or have all 4 processors running but each with a different project?

If the first then set to 25% of multi cpu, if the second then AFAIK it's not possible to enforce - the boinc scheduler decides what and when it will run - so sometimes you will find it running only one project on all 4 processors other times it might be running two, three or four projects - it depends on what the scheduler decides it needs to run.

I was wanting to run one project per Processor (using at least 3 of the 4 cores since I am running 3 projects) along with the GPU. Thank you for answering my question... :-)

Will

That's called processor affinity and is not supported by BOINC. It will not ever be introduced either. You can reach the same sort of thing by just running BOINC, all projects at same resource share.

With BOINC 7.0.25, the REC-scheduler will try to run all projects with the same resource share for the same amount of time, even get the same amount of RAC (when combined with CreditNew). So at first you'll see that each project runs alone on all CPUs, but eventually a shift will happen where the projects are equally divided over the available processor cores.

Note: With mixed CPU and CPU+GPU projects, BOINC will initially get CPU and GPU work from the CPU+GPU project. A fast GPU drives REC so high, that priority for the CPU goes really low. It will eventually run those tasks when they go into EDF, but unless you micromanage not earlier. It will also not get more CPU tasks, because CPU projects have smaller REC and therefore higher priority.

---

That's called processor affinity and is not supported by BOINC. It will not ever be introduced either. You can reach the same sort of thing by just running BOINC, all projects at same resource share.

With BOINC 7.0.25, the REC-scheduler will try to run all projects with the same resource share for the same amount of time, even get the same amount of RAC (when combined with CreditNew). So at first you'll see that each project runs alone on all CPUs, but eventually a shift will happen where the projects are equally divided over the available processor cores.

Note: With mixed CPU and CPU+GPU projects, BOINC will initially get CPU and GPU work from the CPU+GPU project. A fast GPU drives REC so high, that priority for the CPU goes really low. It will eventually run those tasks when they go into EDF, but unless you micromanage not earlier. It will also not get more CPU tasks, because CPU projects have smaller REC and therefore higher priority.

Thank you for the clarificaton.. :-)
Will

[quote]

Are you wanting boinc to only run on 1 processor or have all 4 processors running but each with a different project?

If the first then set to 25% of multi cpu, if the second then AFAIK it's not possible to enforce - the boinc scheduler decides what and when it will run - so sometimes you will find it running only one project on all 4 processors other times it might be running two, three or four projects - it depends on what the scheduler decides it needs to run.

I was wanting to run one project per Processor (using at least 3 of the 4 cores since I am running 3 projects) along with the GPU. Thank you for answering my question... :-)

Will

Maybe assigning one processor core # to a particular task is "CORE AFFINITY", but not wanting BOINC to run one task per project... 4 projects, 4 CPU's, one project per CPU core... simple... It's called running ONE INSTANCE of a project on one core. Please don't confuse the issues even though you answered his question by saying BOINC decides what to do, not the user... which is and continues to be a problem for many folks.

8-)

[quote]

Are you wanting boinc to only run on 1 processor or have all 4 processors running but each with a different project?

If the first then set to 25% of multi cpu, if the second then AFAIK it's not possible to enforce - the boinc scheduler decides what and when it will run - so sometimes you will find it running only one project on all 4 processors other times it might be running two, three or four projects - it depends on what the scheduler decides it needs to run.

I was wanting to run one project per Processor (using at least 3 of the 4 cores since I am running 3 projects) along with the GPU. Thank you for answering my question... :-)

Will

Maybe assigning one processor core # to a particular task is "CORE AFFINITY", but not wanting BOINC to run one task per project... 4 projects, 4 CPU's, one project per CPU core... simple... It's called running ONE INSTANCE of a project on one core. Please don't confuse the issues even though you answered his question by saying BOINC decides what to do, not the user... which is and continues to be a problem for many folks.

8-)

That is a thought I have not entertained yet. Maybe I will try it.... Thank you.... :-)

Will

Maybe assigning one processor core # to a particular task is "CORE AFFINITY", but not wanting BOINC to run one task per project... 4 projects, 4 CPU's, one project per CPU core... simple...

http://en.wikipedia.org/wiki/Processor_affinity writes:

Processor affinity or CPU Pinning. CPU pinning enables mapping and unmapping entire virtual machines or a specific virtual CPU (vCPU), to a physical CPU or a range of CPUs. It is a modification of the native central queue scheduling algorithm in a symmetric multiprocessing operating system. Each task (be it process or thread) in the queue has a tag indicating its preferred / kin processor.

And so 4 projects, 4 CPUs (or 4 cores, or 2 cores + hyper-threaded virtual cores), one project per CPU core == processor affinity. Which is something BOINC doesn't support or will support in the future, since it's proven in tests that there's no gain in doing this.

The closest that you'll get is with BOINC 7.0, if you let it handle things by itself, if you don't overwork your BOINC by setting high minimum + additional buffers, if you let it learn for the first 2 weeks to a month how things work, if the chosen projects all have the same resource share and comparable run-times, if none of the projects goes into HP, etc. etc.

---