Message boards :
Number crunching :
SETI@home v8.22 Windows GPU applications support thread
Message board moderation
Previous · 1 . . . 9 · 10 · 11 · 12 · 13 · 14 · 15 . . . 23 · Next
Author | Message |
---|---|
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
I did not run the suggested steps, but I RESET Project. The new message was: . . This means just what it says, so there will be NO tasks to see. . . This means that if you want to process work on your CPU then you need to re-reun the Lunatics Installer and select the option for CPU processing, the default choice is usually the best option for your CPU hardware. . . These are not tasks, they are image files that form part of the BOINC shell, they display at the appropriate time and place on your screen when running BOINC. . . And this seems to indicated that you did not fully read Zalsters message. After each work request you have to wait 5 mins and 3 secs before you can get any work again. Requesting work manually before then will only result in this message. Stephen . . |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13161 Credit: 1,160,866,277 RAC: 1,873 |
All your old work will show as being done by the old app, but new work will show as correct. Stephen, you do realize what the task is labelled as in your task list or online is meaningless, right? A MB task is a MB task. It could be labelled setiathome_peanut_butter and it would be processed by whatever the MB science application is called in the app_info.xml. As long as you have the r3584 science application in the directory and it is called by the app_info correctly, it doesn't matter how the task is labelled when downloaded. As far as the missing file reference in the MB8_win_x86_SSE3_OpenCL_NV_SoG.aistub file in the mb8_win_x86_sse3_opencl_nv_sog_r3584.7z archive. Here is the .aistub file with the missing reference added in. <app_info> <app> <name>setiathome_v8</name> </app> <file_info> <name>MB8_win_x86_SSE3_OpenCL_NV_SoG_r3584.exe</name> <executable/> </file_info> <file_info> <name>libfftw3f-3-3-4_x86.dll</name> <executable/> </file_info> <file_info> <name>mb_cmdline_win_x86_SSE3_OpenCL_NV.txt</name> </file_info> <app_version> <app_name>setiathome_v8</app_name> <version_num>821</version_num> <platform>windows_intelx86</platform> <avg_ncpus>0.04</avg_ncpus> <max_ncpus>0.2</max_ncpus> <plan_class>opencl_nvidia_SoG</plan_class> <cmdline></cmdline> <coproc> <type>CUDA</type> <count>1</count> </coproc> <file_ref> <file_name>MB8_win_x86_SSE3_OpenCL_NV_SoG_r3584.exe</file_name> <main_program/> </file_ref> <file_ref> <file_name>libfftw3f-3-3-4_x86.dll</file_name> </file_ref> <file_ref> <file_name>MultiBeam_Kernels_r3584.cl</file_name> </file_ref> <file_ref> <file_name>mb_cmdline_win_x86_SSE3_OpenCL_NV.txt</file_name> <open_name>mb_cmdline.txt</open_name> </file_ref> </app_version> </app_info> Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
As I said earlier, you do not need the 8.22 app, I'm quite happy with the 8.20 app and some actually report that the older app is slightly faster than the newer 1. . . Yep you are probably right, v8.20 is probably as good as it gets. But, I am an inquisitive breed of cat and I am vey curious to see what, if any, difference v8.22 makes on the 950. Keith seems to feel that it might make some difference as opposed to more powerful GPU hardware. . . In the end I expect I will re-run Lunatics and go back to doubles. So far all indications are that this was the most productive configuration on this rig. Stephen :) |
Zalster Send message Joined: 27 May 99 Posts: 5517 Credit: 528,817,460 RAC: 242 |
It's probably not going to do anything for a 900 series nvidia card. Raistmer's tweaking of the code was for the really older cards, 500s, 600, 700s Might have included the 400s but I can't recall. Stopped paying attention when i realized it was for cards older than I had. |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13161 Credit: 1,160,866,277 RAC: 1,873 |
. . I stopped BOINC and went through app_info.xml (Keith is right I need to clean it up and create a more efficient version) but I discovered the problem. The download Keith linked me to has a slight discrepancy, it has set the app info to v8.21 .... ( was that a transient version at some stage?) so when I had attempted to rename some files as v8.22 to get r3584 to kick in I sadly joined Miklos and trashed them. So I will now manually change that to v8.22. There is ABSOLUTELY NO NEED to rename any task. EVER. The .aistub in that archive listed the 8.21 app version level for the r3557 science application. That is the last version the Lunatics installer ever installed since that is the last version released since Richard never updated the installer past Beta-06. The r3584 science app got released to Main as the current MB SoG app and was labelled with version 8.22. Again, it makes no difference what the task is labelled with whatever version, it is still the same MB task. What matters is the science application. The current latest one is the r3584 application. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13161 Credit: 1,160,866,277 RAC: 1,873 |
It's probably not going to do anything for a 900 series nvidia card. Raistmer's tweaking of the code was for the really older cards, 500s, 600, 700s Might have included the 400s but I can't recall. Stopped paying attention when i realized it was for cards older than I had. It has more to do with the number of CUDA cores or Compute Units than the card family. I class the 950 on the low end side of compute units at 6. So it MIGHT see a benefit in the r3584 application with regard to system responsiveness. Only testing will tell. I doubt it has anything to do with the speed of the application on tasks unless the adjustment to alleviate lagginess also slows down the task run time. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
It's probably not going to do anything for a 900 series nvidia card. Raistmer's tweaking of the code was for the really older cards, 500s, 600, 700s Might have included the 400s but I can't recall. Stopped paying attention when i realized it was for cards older than I had. . . The old mind over matter thing ... you don't mind so it don't matter ... :) Stephen ps: I would still like to get it working just to see :) :) |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
. . I stopped BOINC and went through app_info.xml (Keith is right I need to clean it up and create a more efficient version) but I discovered the problem. The download Keith linked me to has a slight discrepancy, it has set the app info to v8.21 .... ( was that a transient version at some stage?) so when I had attempted to rename some files as v8.22 to get r3584 to kick in I sadly joined Miklos and trashed them. So I will now manually change that to v8.22. . . I would expect that there should not be a need but hey, horses for courses. As for r3557 that is v8.20, r3584 is v8.22. I am not aware of a v8.21 being released to the public. In any case changing the line in app_info from 821 to 822 didn't fix it, see my message about no kernel files nor a wisdom file. Any ideas? Stephen ? ? |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
. . Actually there is no direct correlation between CUDA cores and CUs. Older card with less CUDA cores can have more CUs, the newer generations seem to have far more cuda cores per CU than their predecessors. But the more CUs the better some things seem to run. Not sure how the apps relate to that difference. Stephen ? |
Zalster Send message Joined: 27 May 99 Posts: 5517 Credit: 528,817,460 RAC: 242 |
So, I guess you download 8.22 application from Mike's World?? You unzipped the file and transferred all the contents into the Seti@home folder? Then used the aistub file? I've never used it but they say it works. For those few machines that I did upgrade to the actual r3584 I manually changed all the values. For those few machines where I only changed the version name, it looks like this. <app_name>setiathome_v8</app_name> <version_num>822</version_num> <platform>windows_intelx86</platform> <api_version>7.5.0</api_version> <avg_ncpus>0.04</avg_ncpus> <max_ncpus>0.2</max_ncpus> <plan_class>opencl_nvidia_SoG</plan_class> <cmdline></cmdline> <coproc> <type>CUDA</type> <count>1</count> </coproc> <file_ref> <file_name>MB8_win_x86_SSE3_OpenCL_NV_SoG_r3557.exe</file_name> <main_program/> </file_ref> <file_ref> <file_name>libfftw3f-3-3-4_x86.dll</file_name> </file_ref> <file_ref> <file_name>MultiBeam_Kernels_r3557.cl</file_name> </file_ref> <file_ref> <file_name>mb_cmdline_win_x86_SSE3_OpenCL_NV_SoG.txt</file_name> <open_name>mb_cmdline.txt</open_name> </file_ref> </app_version> If you manually install, you will need to change 8.20 to 8.22 and 3557 to 3584 |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13161 Credit: 1,160,866,277 RAC: 1,873 |
I believe his problem came from the mb8_win_x86_sse3_opencl_nv_sog_r3584.7z downloaded from Mike's World. That file doesn't have the complete MB8_win_x86_SSE3_OpenCL_NV_SoG.aistub file. It is missing the: <file_ref> <file_name>MultiBeam_Kernels_r3584.cl</file_name> </file_ref> entry in the .aistub file for some reason. That is why I posted the corrected .aistub file contents in my earlier post to him. If he had simply copy and pasted the text into the existing .aistub file and reran aimerge, the app_info would generate the wisdom files. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Miklos M. Send message Joined: 5 May 99 Posts: 955 Credit: 136,115,648 RAC: 73 |
Thank you everyone for your help. It is running like a champ again with 8.22's. Reinstalled and went to Mike's. |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
. . Thanks Zalster, yes it was the download from Mike's World and yes I copied the files into the project directory and ran aimerge.com. I will add the file reference for the .cl kernel file. Stephen . . |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
I believe his problem came from the mb8_win_x86_sse3_opencl_nv_sog_r3584.7z downloaded from Mike's World. That file doesn't have the complete MB8_win_x86_SSE3_OpenCL_NV_SoG.aistub file. It is missing the: . . Thanks Keith, . . I don't know why/how but I must have missed that post, I'll scroll back and have a look. Stephen . . |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
. . OK I found it ... . . Thanks for the reply. . . Maybe I have it wrong but I thought the schedulers added the version number to the file name according to the app requesting work, which is why files with a version number that doesn't match an installed app get trashed. That is to say a host running r3584 requesting work will get v822 named files, and so the app_info has to match the version number the scheduler expects it to. Stephen . . |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
. . OK, here's the thing ... . . I added the lines you provided into the aistub and ran aimerge. App_info now has that info but it still is running r3557, so downloads are still v8.20. Do I need to remove the references to r3557 or change them to r3584 to get it to run ? . . This is why I like Lunatics :) Stephen ? ? |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13161 Credit: 1,160,866,277 RAC: 1,873 |
Stephen, the app_info is what is used to call the science app. It it calls the r3557 app with version 8.20, that is what any downloaded work will be labelled. And it will process with the r3557 app irregardless if you have the r3584 app in the directory. BOINC only does what you tell it to do with the app_info. It can't decide on its own just to run the r3584 app just because you put it there in the directory and wish for it. You have to edit the app_info that is created by Lunatics and replace all instances of the r3557 application in the file with the r3584 app. It you want to label the tasks with the 8.22 version label, you will have to replace all instances of 8.20 with 8.22. Notepad Find function with Find and Replace All works just fine in doing that. You will also have to replace all instances of the r3557.CL file with the r3584.CL file also. I'm not sure why you can't get this concept of replacing what you don't want with what you want in app_info. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13161 Credit: 1,160,866,277 RAC: 1,873 |
Stephen, just to make it simple for you. Here is the app_info with the r3584 app and the Lunatics cpu AP and MB applications. <app_info> <app> <name>astropulse_v7</name> </app> <file_info> <name>AP7_win_x64_AVX_CPU_r2692.exe</name> <executable/> </file_info> <file_info> <name>libfftw3f-3-3-4_x64.dll</name> <executable/> </file_info> <file_info> <name>ap_cmdline_win_x64_AVX_CPU.txt</name> </file_info> <app_version> <app_name>astropulse_v7</app_name> <version_num>703</version_num> <platform>windows_x86_64</platform> <plan_class>sse2</plan_class> <cmdline></cmdline> <file_ref> <file_name>AP7_win_x64_AVX_CPU_r2692.exe</file_name> <main_program/> </file_ref> <file_ref> <file_name>libfftw3f-3-3-4_x64.dll</file_name> </file_ref> <file_ref> <file_name>ap_cmdline_win_x64_AVX_CPU.txt</file_name> <open_name>ap_cmdline.txt</open_name> </file_ref> </app_version> <name>astropulse_v7</name> <file_info> <name>AP7_win_x86_SSE2_OpenCL_NV_r2887.exe</name> <executable/> </file_info> <file_info> <name>libfftw3f-3-3-4_x86.dll</name> <executable/> </file_info> <file_info> <name>AstroPulse_Kernels_r2887.cl</name> </file_info> <file_info> <name>ap_cmdline_win_x86_SSE2_OpenCL_NV.txt</name> </file_info> <app_version> <app_name>astropulse_v7</app_name> <version_num>710</version_num> <platform>windows_x86_64</platform> <avg_ncpus>0.04</avg_ncpus> <max_ncpus>0.2</max_ncpus> <plan_class>opencl_nvidia_100</plan_class> <cmdline></cmdline> <coproc> <type>CUDA</type> <count>1</count> </coproc> <file_ref> <file_name>AP7_win_x86_SSE2_OpenCL_NV_r2887.exe</file_name> <main_program/> </file_ref> <file_ref> <file_name>libfftw3f-3-3-4_x86.dll</file_name> </file_ref> <file_ref> <file_name>AstroPulse_Kernels_r2887.cl</file_name> </file_ref> <file_ref> <file_name>ap_cmdline_win_x86_SSE2_OpenCL_NV.txt</file_name> <open_name>ap_cmdline.txt</open_name> </file_ref> </app_version> <app> <name>setiathome_v8</name> </app> <file_info> <name>MB8_win_x64_AVX_VS2010_r3330.exe</name> <executable/> </file_info> <file_info> <name>libfftw3f-3-3-4_x64.dll</name> <executable/> </file_info> <file_info> <name>mb_cmdline_win_x64_AVX_VS2010.txt</name> </file_info> <app_version> <app_name>setiathome_v8</app_name> <version_num>800</version_num> <platform>windows_x86_64</platform> <api_version>7.5.0</api_version> <file_ref> <file_name>MB8_win_x64_AVX_VS2010_r3330.exe</file_name> <main_program/> </file_ref> <file_ref> <file_name>libfftw3f-3-3-4_x64.dll</file_name> </file_ref> <file_ref> <file_name>mb_cmdline_win_x64_AVX_VS2010.txt</file_name> <open_name>mb_cmdline.txt</open_name> </file_ref> </app_version> <app> <name>setiathome_v8</name> </app> <file_info> <name>MB8_win_x86_SSE3_OpenCL_NV_SoG_r3584.exe</name> <executable/> </file_info> <file_info> <name>libfftw3f-3-3-4_x86.dll</name> <executable/> </file_info> <file_info> <name>MultiBeam_Kernels_r3584.cl</name> </file_info> <file_info> <name>mb_cmdline_win_x86_SSE3_OpenCL_NV_SoG.txt</name> </file_info> <app_version> <app_name>setiathome_v8</app_name> <version_num>822</version_num> <platform>windows_x86_64</platform> <api_version>7.5.0</api_version> <avg_ncpus>0.04</avg_ncpus> <max_ncpus>0.2</max_ncpus> <plan_class>opencl_nvidia_SoG</plan_class> <cmdline></cmdline> <coproc> <type>CUDA</type> <count>1</count> </coproc> <file_ref> <file_name>MB8_win_x86_SSE3_OpenCL_NV_SoG_r3584.exe</file_name> <main_program/> </file_ref> <file_ref> <file_name>libfftw3f-3-3-4_x86.dll</file_name> </file_ref> <file_ref> <file_name>MultiBeam_Kernels_r3584.cl</file_name> </file_ref> <file_ref> <file_name>mb_cmdline_win_x86_SSE3_OpenCL_NV_SoG.txt</file_name> <open_name>mb_cmdline.txt</open_name> </file_ref> </app_version> </app_info> Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Stephen "Heretic" Send message Joined: 20 Sep 12 Posts: 5557 Credit: 192,787,363 RAC: 628 |
Stephen, the app_info is what is used to call the science app. It it calls the r3557 app with version 8.20, that is what any downloaded work will be labelled. And it will process with the r3557 app irregardless if you have the r3584 app in the directory. BOINC only does what you tell it to do with the app_info. It can't decide on its own just to run the r3584 app just because you put it there in the directory and wish for it. . . OK here's the thing ... . . I didn't just do that, I had already added the r3584 info to the app_info file. . . First I was told, by yourself, to edit the file as you have just re-described but then I was told just use the aistub and run aimerge. Believing that would work I did just that, but hey, it doesn't work. I am now realising how much work Richards puts into getting the Lunatics installer right. So I am back to square one and facing the full manual edit. The fact that the aistub included in the download is deficient didn't help. Stephen [edit] That was bad advice, the only app entry that should say 822 is the one for the new app itself, the old entries should have been left as they were so the existing files would not get ghosted as they just have been. That is why Richard has so many entries and I knew better than to do that but I was mindlessly just following instructions ... D'oh! <shrug> |
Wiggo Send message Joined: 24 Jan 00 Posts: 34744 Credit: 261,360,520 RAC: 489 |
Ok, run your GPU cache down, then move the old r3557 aistub to the oldApp_backup folder, edit the newer aistub to reflect the correct entries and then run the aimerge.cmd again, that should fix it. All I can say is that it all worked perfectly for me back in the changeover from MB V6 to MB V7 days. Cheers. |
©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.