Posts by captainjack

1) Message boards : Number crunching : Python apps for GPU hosts 4.03 (cuda1131) using a LOT of CPU (Message 59344)
Posted 26 Sep 2022 by captainjack
Post:
Keith,

I hope I am understanding your question correctly,

Have you set the <use_all_gpus> flag in the cc_config.xml file to "1"?

Requires a restart if the flag needs to be turned on.
2) Message boards : Server and website : python tasks get to 2.00% and hang (Message 58903)
Posted 11 Jun 2022 by captainjack
Post:
AFAIK, GPUGRID does not use a VirtualBox VM. At lease it doesn't on either of my machines.

Remaining estimate always starts out really high until several tasks are completed successfully, then it will start being more accurate.

A Python task will use about 65% of available CPU threads. For example, on my 6 core 12 thread CPU, Python tasks will use 7-8 threads.

On all of your tasks that I checked, it shows an error message of "OSError: [WinError 1455] The paging file is too small for this operation to complete. Error loading "C:\ProgramData\BOINC\slots\13\lib\site-packages\torch\lib\cudnn_cnn_infer64_8.dll" or one of its dependencies.
Traceback (most recent call last):"

Suggest that you increase the size of your paging file, run a python task by itself and watch system usage, run the python task through to completion without interruption.

Let us know if that helps.
3) Message boards : News : Experimental Python tasks (beta) - task description (Message 58878)
Posted 29 May 2022 by captainjack
Post:
Thank you abouh for responding,

I looked through my saved messages from the task to see if there was anything else I could find that might be of value and couldn't find anything.

In regard to the "out of memory" error, I tried to read through the stackoverflow link about the memory error. It is way above my level of technical expertise at this point, but it seemed like the amount of nvidia memory might have something to do with it. I am using an antique GTX970 card. It's old but still works.

Good luck coming up with a solution. If you want me to do any more testing, please let me know.
4) Message boards : News : Experimental Python tasks (beta) - task description (Message 58871)
Posted 28 May 2022 by captainjack
Post:
Windows 10 machine running task 32899765. Had a power outage. When the power came back on, task was restarted but just sat there doing nothing. The stderr.txt file showed the following error:

file pythongpu_windows_x86_64__cuda102.tar
already exists. Overwrite with
pythongpu_windows_x86_64__cuda102.tar?
(Y)es / (N)o / (A)lways / (S)kip all / A(u)to rename all / (Q)uit?


Task was stalled waiting on a response.

BOINC was stopped and the pythongpu_windows_x86_64__cuda102.tar file was removed from the slots folder.

Computer was restarted then the task was restarted. Then the following error message appeared several times in the stderr.txt file.

OSError: [WinError 1455] The paging file is too small for this operation to complete. Error loading "C:\ProgramData\BOINC\slots\0\lib\site-packages\torch\lib\caffe2_detectron_ops_gpu.dll" or one of its dependencies.
Detected memory leaks!


Page file size was increased to 64000MB and rebooted.

Started task again and still got the error message about page file size too small. Then task abended.

If you need more info about this task, please let me know.
5) Message boards : Number crunching : GPU Idling, progress halted at 2% (Message 58800)
Posted 11 May 2022 by captainjack
Post:
jjch asked:
Do I need to reserve more CPU than the default 0.988 for these tasks?


Yes. Try running the task by itself to see how many CPUs it uses. On my 8 core 16 thread machine, it will use about 10 threads. On my 6 core 12 thread machine, it will use about 8 threads.
6) Message boards : News : Experimental Python tasks (beta) - task description (Message 58783)
Posted 4 May 2022 by captainjack
Post:
abouh asked
Right, I wish there was a way to specify that to BOINC on our side... does adjusting the app_config.xml help? I guess that has to be done of the user side


I tried that, but boinc manager on my pc will overallocate CPU's. I am currently running multicore atlas cpu tasks from lhc alongside the python tasks from gpugrid. The atlas tasks are set to use 8 CPU's and the python tasks are set to use 10 CPU's. The example for this response is on an AMD cpu with 8 cores/16 threads. BOINC is set to use 15 threads. It will run one gpugrid python 10 thread task and one lhc 8 thread task at the same time. That is 18 threads running on a 15 thread cpu.

Here is my app_config for gpugrid:

<app_config>
   <app>
      <name>acemd3</name>
        <gpu_versions>
           <gpu_usage>1</gpu_usage>
           <cpu_usage>1</cpu_usage>
        </gpu_versions>
   </app>
   <app>
      <name>PythonGPU</name>
        <cpu_usage>10</cpu_usage>
        <gpu_versions>
           <gpu_usage>1</gpu_usage>
           <cpu_usage>10</cpu_usage>
        </gpu_versions>
        <app_version>
           <app_name>PythonGPU</app_name>
           <plan_class>cuda1121</plan_class>
           <avg_ncpus>10</avg_ncpus>
           <ngpus>1</ngpus>
           <cmdline>--nthreads 10</cmdline>
        </app_version>
   </app>

   <app>
      <name>PythonGPUbeta</name>
        <cpu_usage>10</cpu_usage>
        <gpu_versions>
           <gpu_usage>1</gpu_usage>
           <cpu_usage>10</cpu_usage>
        </gpu_versions>
        <app_version>
           <app_name>PythonGPU</app_name>
           <plan_class>cuda1121</plan_class>
           <avg_ncpus>10</avg_ncpus>
           <ngpus>1</ngpus>
           <cmdline>--nthreads 10</cmdline>
        </app_version>
   </app>

   <app>
      <name>Python</name>
        <cpu_usage>10</cpu_usage>
        <gpu_versions>
           <gpu_usage>1</gpu_usage>
           <cpu_usage>10</cpu_usage>
        </gpu_versions>
        <app_version>
           <app_name>PythonGPU</app_name>
           <plan_class>cuda1121</plan_class>
           <avg_ncpus>10</avg_ncpus>
           <ngpus>1</ngpus>
           <cmdline>--nthreads 10</cmdline>
        </app_version>
   </app>

   <app>
      <name>acemd4</name>
        <gpu_versions>
           <gpu_usage>1</gpu_usage>
           <cpu_usage>1</cpu_usage>
        </gpu_versions>
   </app>
</app_config>


And here is my app_config for lhc:

<app_config>
  <app>
      <name>ATLAS</name>
        <cpu_usage>8</cpu_usage>
  </app>
  <app_version>
      <app_name>ATLAS</app_name>
        <plan_class>vbox64_mt_mcore_atlas</plan_class>
           <avg_ncpus>8</avg_ncpus>
           <cmdline>--nthreads 8</cmdline>
  </app_version>
</app_config>


If anyone has any suggestions for changes to the app_config files, please let me know.
7) Message boards : News : ACEMD 4 (Message 58763)
Posted 28 Apr 2022 by captainjack
Post:
ACEMD 4 tasks now seem to be processing ok on my antique GTX 970. I had previously reported that they were all failing.
Two GAFF2 tasks completed and validated. 1 QM task underway for 25 minutes and processing normally.
8) Message boards : News : ACEMD 4 (Message 58717)
Posted 25 Apr 2022 by captainjack
Post:
Received 25 tasks on Linux. Ran about 45 - 60 seconds. All error. Error message:
process exited with code 195 (0xc3, -61)


Received 0 tasks on Windows.
9) Message boards : News : ACEMD 4 (Message 58645)
Posted 14 Apr 2022 by captainjack
Post:
Thu 14 Apr 2022 09:27:26 AM CDT | GPUGRID | Aborting task T1_GAFF2_frag_00-RAIMIS_NNPMM-0-1-RND6653_1: exceeded elapsed time limit 7231.33 (1000000.00G/138.29G)
Thu 14 Apr 2022 09:27:28 AM CDT | GPUGRID | Computation for task T1_GAFF2_frag_00-RAIMIS_NNPMM-0-1-RND6653_1 finished
Thu 14 Apr 2022 09:27:28 AM CDT | GPUGRID | Output file T1_GAFF2_frag_00-RAIMIS_NNPMM-0-1-RND6653_1_4 for task T1_GAFF2_frag_00-RAIMIS_NNPMM-0-1-RND6653_1 exceeds size limit.
Thu 14 Apr 2022 09:27:28 AM CDT | GPUGRID | File size: 137187308.000000 bytes. Limit: 10000000.000000 bytes
10) Message boards : News : Experimental Python tasks (beta) - task description (Message 58588)
Posted 31 Mar 2022 by captainjack
Post:
These tasks seem to run much better on my machines if I allocate 6 CPU's (threads) to each task. I managed to run one by itself and watched the performance monitor for CPU usage. During the initiation phase (about 5 minutes), the task used ~6 CPU's (threads). After the initiation phase, the CPU usage was in an oscillating pattern that was between ~2 and ~5 threads. Task ran very quickly and has been validated. Please let me know if you have questions.
11) Message boards : Server and website : Extremely slow download rate (Message 58412)
Posted 2 Mar 2022 by captainjack
Post:
I've got one of those too. Mine is downloading at 20.42 KBps. After 4:41 elapsed time it has only downloaded .33GB out of a total download size of 2.86GB. Deadline is 24 hours after the task arrived. No way it will finish the download within the 24 hour deadline.
12) Message boards : Graphics cards (GPUs) : Two different GPUs, 1, only 1 GPU getting tasks (Message 54266)
Posted 6 Apr 2020 by captainjack
Post:
Elevator Startups,

Your startup log shows that your second gpu is ignored by config. You probably need to add the <use_all_gpus>1</use_all_gpus> statement to your cc_config.xml file.
13) Message boards : Number crunching : have a lot of stuck tasks, abort some? (Message 53155)
Posted 27 Nov 2019 by captainjack
Post:
JStateson wrote:

I suspect the first 67mb files caused a problem which was compouned by subsequent ones of same size all trying to upload concurrently. Need to figure a was to stop this.


Just a thought, in the cc_config.xml file there is an option for 
<max_file_xfers_per_project>N</max_file_xfers_per_project>
.
Maybe that would help.
14) Message boards : GPUGRID CAFE : With high-end Intel GPUs coming: Maybe OpenCl is the way to go ahead? (Message 53097)
Posted 24 Nov 2019 by captainjack
Post:
Carl Philip said:

So I still don't understand why my R7 is 8 times faster in Milkyway compared to the GT730.


From what I remember, Milkyway performs double precision calculations on the GPU. A GPU that supports double precision natively will be much faster than one that does not. Depending on the particular model, your R7 is much faster at double precision math that your GT730 according to the ratings on Wikipedia.

I used to run Milkyway on an AMD 6950 and it would run circles around my GT970's.
15) Message boards : Multicore CPUs : QC tests for windows (Message 51661)
Posted 28 Mar 2019 by captainjack
Post:
Did a project reset and started three more tasks. Two task finished successfully and the other aborted with:

3/28/2019 10:37:48 AM | GPUGRID | Aborting task 170r5-TONI_TST11-0-1-RND6393_2: exceeded disk limit: 34955.30MB > 28610.23MB


As a separate test, I tried the tasks on a different Windows PC number 489422. All four tasks aborted with the following message:

<core_client_version>7.14.2</core_client_version>
<![CDATA[
<message>
(unknown error) - exit code 195 (0xc3)</message>
<stderr_txt>
10:28:17 (15156): wrapper (7.9.26016): starting
10:28:17 (15156): wrapper: running install_miniconda.bat ()
10:28:18 (15156): install_miniconda.bat exited; CPU time 0.000000
10:28:18 (15156): app exit status: 0x1
10:28:18 (15156): called boinc_finish(195)


I did a project reset and got the same result. Am I missing something?

16) Message boards : Multicore CPUs : QC tests for windows (Message 51658)
Posted 28 Mar 2019 by captainjack
Post:
And another:

3/28/2019 9:07:38 AM | GPUGRID | Aborting task 1445r7-TONI_TST11-0-1-RND1466_0: exceeded disk limit: 44800.51MB > 28610.23MB

Please let us know when this is fixed and I will start up some more test tasks.
17) Message boards : Graphics cards (GPUs) : 2 GPU, not the same % of utilization (Message 51657)
Posted 28 Mar 2019 by captainjack
Post:
What is the Maximum PCIe Link Width and Maximum PCIe Link Speed of each of the two slots where the GPUs are inserted?
18) Message boards : Multicore CPUs : QC tests for windows (Message 51653)
Posted 28 Mar 2019 by captainjack
Post:
Got this error message on the current batch of QC test for Windows:

3/28/2019 7:47:29 AM | GPUGRID | Aborting task 1445r30-TONI_TST11-0-1-RND5643_0: exceeded disk limit: 44800.51MB > 28610.23MB
19) Message boards : News : CPU jobs on Linux (Message 51518)
Posted 17 Feb 2019 by captainjack
Post:
biodoc asked:

Has anyone got this working with a linux VM on windows?

Yes, I have two machines set up that I use occasionally, 489909 and 490622. Host O/S is Windows 10 Home, guest O/S is Ubuntu 18.10.
20) Message boards : Multicore CPUs : New QC app (Message 51298)
Posted 12 Jan 2019 by captainjack
Post:
STEVE,

I prefer this line of thinking, general business users usually run Windows, graphics artist and publishers usually run Macs, and scientists usually run Linux. Almost all of the top 500 super-computers in the world run on Linux. I'm guessing that most of the scientific research applications natively run on Linux. It was easier for the GPUGRID scientist to get the QC CPU app running on Linux PC's so that was the initial release.

They tried to get the QC CPU to run on Windows a few months ago using the Windows Subsystem for Linux (WSL) but ran into problems with inconsistency between Windows releases. This is their next attempt to get the QC app to run on Windows. Looks like this one might work when they get the parameters set up for the size of tasks involved.


Next 20

©2026 Universitat Pompeu Fabra