boinc_7.0.36_windows_x86_64.exe
Printable View
boinc_7.0.36_windows_x86_64.exe
From knreed over at the WCG forum.
Quote:
So here is the problem:
17/11/2012 02:58:23 | World Community Grid | [error] App version returned from anonymous platform project; ignoring
17/11/2012 02:58:23 | World Community Grid | [error] No app version found for app hcc1 platform windows_intelx86 ver 656 class ati_hcc1; discarding X0960073631347200608011011_4
A job gets assigned that uses the old app version.
The host is anonymous platform, so it ignores the app version sent.
The host can't find app version that matches the platform, version num and plan_class so it discards the job
The problem is that discarding the job does not report to the server that the client isn't running it. Thus the job is still assigned to you.
Next request to the server, you get sent the job again. This continues.
Even worse, each time the job is sent to you the deadline for the job is re-evaluated and possibly slightly increased. Thus it can potentially never pass its deadline.
When I started digging into this problem today, there were a lot of computers who were repeatedly being resent the same jobs.
This issue occurs when all three of app_info.xml is used, homogenous_app_version and resend_results.
Resolving this issue for the long run is going to be somewhat tricky. As a result, what I am doing now is changing the app_version on the workunits to all be at the 705 level. The new binaries are backward compatible so this shouldn't be an issue. This should return life to normal for now.
Thanks for the above clarification Nanoprobe.
Are you still rercommending BOINC app manager V 7.0.36 beta or is 7.0.38 the one to go for now?
Hello again nanoprobe,
Thanks for quick reply. I have now uodated BOINC manager to V7.0.36 from V7.0.28. Transition was smooth and painless. Here is a couple more questions for you:
1- Do you still recommend using the dual 6.56 / 7.05 hcc app_info file that you posted earlier in light of of knreed modding the old 6.56 hcc to 7.05 hcc?
2. In BOINC Manager "preferences/Disk & memory usage/memory usage" What % settings do you recommend for the "comp in use" and "at idle" settings? (Default appears to be 50% and 75%)
Best regards,
Jean-Guy
Yes, I'm still using it as posted. I'm of the mind if it's working, don't mess with it.
That's really up to what works best for you. On my dedicated crunches I have them both set @ 90% but they are 64 bit OS with 8 gigs of ram. On my everyday use machine (this one) I use 75% and 85% because this is an XP machine with only 4 gigs of memory which XP can't use all of anyway because it's 32 bit. There are some of the WCG CPU projects that are pretty memory intensive(forget which ones) and I would get the occasional waiting for memory on the XP box running 8 WUs at a time but since I'm only running the GPU app and limited to running only a single task memory is not an issue.
So I am creating a new app_info for DSFL. I am still unsure on how to find this info for the project:
Code:<version_num>625</version_num>
<platform>windows_x86_64</platform>
<avg_ncpus>1.0</avg_ncpus>
<max_ncpus>1.0</max_ncpus>
<flops>3347548492.458962</flops>
<api_version>7.1.0</api_version>
Full projected app_info [minus the HCC stuff]:
Code:<app>
<name>dsfl</name>
<user_friendly_name>Drug Search for Leishmaniasis</user_friendly_name>
</app>
<file_info>
<name>wcgrid_dsfl_vina_6.25_windows_x86_64</name>
<executable/>
</file_info>
<file_info>
<name>wcgrid_dsfl_vina_prod_x86_64.exe.6.25</name>
<executable/>
</file_info>
<file_info>
<name>wcgrid_dsfl_gfx_prod_x86_64.exe.6.25</name>
<executable/>
</file_info>
<app_version>
<app_name>dsfl</app_name>
<version_num>625</version_num>
<platform>windows_x86_64</platform>
<avg_ncpus>1.0</avg_ncpus>
<max_ncpus>1.0</max_ncpus>
<flops>3347548492.458962</flops>
<api_version>7.1.0</api_version>
<file_ref>
<file_name>wcgrid_dsfl_vina_6.25_windows_x86_64</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>wcgrid_dsfl_vina_prod_x86_64.exe.6.25</file_name>
<open_name>AutoDockVina64.exe</open_name>
</file_ref>
<file_ref>
<file_name>wcgrid_dsfl_gfx_prod_x86_64.exe.6.25</file_name>
<open_name>graphics_app</open_name>
</file_ref>
</app_version>
<app_version>
<app_name>dsfl</app_name>
<version_num>625</version_num>
<platform>windows_intelx86</platform>
<avg_ncpus>1.000000</avg_ncpus>
<max_ncpus>1.000000</max_ncpus>
<flops>3347548492.458962</flops>
<api_version>7.1.0</api_version>
<file_ref>
<file_name>wcgrid_dsfl_vina_6.25_windows_intelx86</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>wcgrid_dsfl_vina_prod_x86.exe.6.25</file_name>
<open_name>AutoDockVina64.exe</open_name>
</file_ref>
<file_ref>
<file_name>wcgrid_dsfl_gfx_prod_x86.exe.6.25</file_name>
<open_name>graphics_app</open_name>
</file_ref>
</app_version>
jspace gave me this:The following may help too:Quote:
Look for the client_state.xml file. D_A mentioned it somewhere.
It has the info for <app_version> and the name. You will need to edit a couple lines, specifically in the section that defines which files need to be present. In client_state it just says <file>, but it needs to say <file_info>.
I just kind of guess what needs to be in there and see if it works:shrug:
http://boinc.berkeley.edu/wiki/Anonymous_platform
http://boincfaq.mundayweb.com/index....499&language=1
and
http://www.broadbandreports.com/r0/d...a/App_Info.pdf
So, the way I see it is this: Get your selected project running on rig without any app_info. go to your client state file and print it out maybe. using the above info try to make the conversions.
run the app_info you built and see if it works. repeat until it does :D
Is it possible to address more tasks/CPU core with app_info settings? I tried taking the SN2S and edited it for AMD and changed average nCPUs inside the SN2S code to 0.5 and still 4 threads total run on this CPU. Is there a way to change this with app_info?
Thanks. I didn't think that was possible. Oh well. I did assign 16 GPU tasks to my 7950 so that should give me a few extra hours of runtime.
How do I get the cpu tasks to run at the same time as the gpu tasks? Currently running 4 gpu units on 2 cores (.5cpu+.25gpu) and folding on the other two. I have cpu usage set to 100%, HHC and search for leish as selected projects, but with the app_info.xml I only get HCC units. Would like to switch the 2 cores from f@h to BOINC.
Thanks,
Yeroon
Change how many cpu threads BOINC allows:
cc_config.xml
<ncpus>N</ncpus>
http://boinc.berkeley.edu/wiki/Cc_config.xml
Quote:
Act as if there were N CPUs; i.e. to simulate 2 CPUs on a machine that has only 1. To use the number of available CPUs, set the value to -1 (was 0 which in newer clients really means zero to e.g. only allow GPU computing).
rig running dry with this error message :shrug:
Code:45 World Community Grid 20.11.2012 21:05:48 Requesting new tasks for ATI
46 World Community Grid 20.11.2012 21:05:53 Scheduler request completed: got 0 new tasks
47 World Community Grid 20.11.2012 21:05:53 No tasks sent
48 World Community Grid 20.11.2012 21:05:53 Tasks are committed to other platforms
They are working on it:
http://www.xtremesystems.org/forums/...app_info-files
Once I get HCC stable in the GPU, I'll get the DSFL one going as I'm crunching it now.
Evantaur:
This pause in GPU work has allowed me to test the coloured app_info in the first post.
I removed the nVidia info (I have Ati) and the dividing lines between each project then ran it as my rig ran out of work.
So, now I have set GPU to run 8 threads on 6 cores leaving 2 cores to sweep up the cpu work later.
As my GPU work ran out BOINC just started to allow cpu work on those cores too.
About the only thing I am unable to do is set GPU to download a full 400 units when the work becomes available again.
Anyone had any joy yet with the remaining <info's> (HFCC, HCMD, DSFL)??? If not I will likely start on that tomorrow.
EDIT: Just to clarify, I did the following
Tidy the coloured app_info from the first post as mentioned above then hold in notepad
Remove the existing app_info.
Change my "home" profile to include all except HFCC, HCMD, DSFL and then allow cpu work then put the rig on this profile. Make sure I only have 0.5 days cache set both in profile and on rig
Stop BOINC then re-start in normal running mode whereupon it downloaded a bunch of cpu wu's (0.5 days worth).
Allow the rig to work these first new wu's for a few minutes (about 10)
Copy the file held in notepad into worldcommunitygrid. org folder as app_info.xml
Pause and then exit BOINC.
Re-start Boinc and "run always"
Update and questions.
I'm waiting on the CPU from Ontrac delivery.
The syringe of thermal paste had a air bubble and part was dry so I'm overnighting 2 tubes to go with the water cooler for the cpu.
Nanoprobe said to turn off the CrossFire when not gaming... is that correct? Does that mean I need 2 app_info files? or 2 boinc programs running at the same time for the two 7950 video cards or what?
Also in a post I read "And they did warn us that they could not guarantee that GPUs (specifically AMD GPUs) would be used in future projects..."
http://www.xtremesystems.org/forums/...upply-of-tasks
2nd page
Author Otis11
Can anybody expound ... is it because OpenCL 1.0 is in AMD card hardware and not 1.1?
Bruce
do any app_info files work for AMD APU's?
I would recommend you turn crossfire off when you're crunching. If you game you can turn it back on for that. You will need only 1 app_info file and 1 boinc program running for both cards. Some people have a problem with boinc seeing both cards. If that happens to you there is a simple fix but we'll wait to see if you have the problem before posting the fix.
WCG has made it clear that there may or may not be additional GPU based projects after HCC is finished. We'll just have to wait and see. I personally don't think their decision will be based on which platform any future apps may use. Its really up to the project scientists and developers, not WCG.
Here is my working DSFL:
Code:<app>
<name>dsfl</name>
<user_friendly_name>Drug Search for Leishmaniasis</user_friendly_name>
</app>
<file_info>
<name>wcgrid_dsfl_vina_6.25_windows_x86_64</name>
<executable/>
</file_info>
<file_info>
<name>wcgrid_dsfl_vina_prod_x86_64.exe.6.25</name>
<executable/>
</file_info>
<file_info>
<name>wcgrid_dsfl_gfx_prod_x86_64.exe.6.25</name>
<executable/>
</file_info>
<app_version>
<app_name>dsfl</app_name>
<version_num>625</version_num>
<platform>windows_x86_64</platform>
<avg_ncpus>1</avg_ncpus>
<max_ncpus>.125</max_ncpus>
<api_version>6.25</api_version>
<file_ref>
<file_name>wcgrid_dsfl_vina_6.25_windows_x86_64</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>wcgrid_dsfl_vina_prod_x86_64.exe.6.25</file_name>
<open_name>AutoDockVina64.exe</open_name>
</file_ref>
<file_ref>
<file_name>wcgrid_dsfl_gfx_prod_x86_64.exe.6.25</file_name>
<open_name>graphics_app</open_name>
</file_ref>
</app_version>
Hello Retro77,
1- The last portion of your working DSFL app_info file appears to be missing. Can you please provide the full version. (no window slider present..?)
2- Why are you using an <avg ncpus> of 1 and an <max ncpus> of 0.125? (just curious..)
Thanks,
Jean-Guy