Astropulse 7,00 released for Linux 32&64, Win 32&64, Win32+AMD/NVIDIA/Intel GPU

Message boards : News : Astropulse 7,00 released for Linux 32&64, Win 32&64, Win32+AMD/NVIDIA/Intel GPU
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 32 · 33 · 34 · 35

AuthorMessage
Profile Raistmer
Volunteer tester
Avatar

Send message
Joined: 18 Aug 05
Posts: 2423
Credit: 15,878,738
RAC: 0
Russia
Message 54129 - Posted: 30 Apr 2015, 6:46:40 UTC
Last modified: 30 Apr 2015, 6:47:45 UTC

And we have few undeployed Linux AP apps still. Linux part lags behind and peoples on main get worry. But w/o at least short beta they can't get new apps...
News about SETI opt app releases: https://twitter.com/Raistmer
ID: 54129 · Report as offensive
Urs Echternacht
Volunteer tester
Avatar

Send message
Joined: 18 Jan 06
Posts: 1038
Credit: 18,734,730
RAC: 0
Germany
Message 54139 - Posted: 2 May 2015, 21:36:41 UTC - in response to Message 54129.  

And we have few undeployed Linux AP apps still. Linux part lags behind and peoples on main get worry. But w/o at least short beta they can't get new apps...

Additional there was one plan class for Mac OS X left out on the last actualization on main :
Mac OS X/64-bit Intel 7.04 (opencl_ati_lion) 7 Oct 2014, 21:10:29 UTC 0 GigaFLOPS
That plan class should also deploy 7.07 apps (same app as opencl_ati_mac plan class), because it represents only a split not a separate app.
_\|/_
U r s
ID: 54139 · Report as offensive
Profile Raistmer
Volunteer tester
Avatar

Send message
Joined: 18 Aug 05
Posts: 2423
Credit: 15,878,738
RAC: 0
Russia
Message 54152 - Posted: 5 May 2015, 21:42:42 UTC

We have new Linux AP builds deployed, please test.
News about SETI opt app releases: https://twitter.com/Raistmer
ID: 54152 · Report as offensive
Urs Echternacht
Volunteer tester
Avatar

Send message
Joined: 18 Jan 06
Posts: 1038
Credit: 18,734,730
RAC: 0
Germany
Message 54163 - Posted: 8 May 2015, 0:23:25 UTC - in response to Message 54152.  

We have new Linux AP builds deployed, please test.

Looking through the result list of this host i can resume that the "bool2"-issue on NV-GPUs with newest drivers (newer than 347.xx) is solved.
_\|/_
U r s
ID: 54163 · Report as offensive
Profile Raistmer
Volunteer tester
Avatar

Send message
Joined: 18 Aug 05
Posts: 2423
Credit: 15,878,738
RAC: 0
Russia
Message 54164 - Posted: 8 May 2015, 11:12:40 UTC - in response to Message 54163.  

We have new Linux AP builds deployed, please test.

Looking through the result list of this host i can resume that the "bool2"-issue on NV-GPUs with newest drivers (newer than 347.xx) is solved.

Then we can proceed to release on main for these builds - right?
News about SETI opt app releases: https://twitter.com/Raistmer
ID: 54164 · Report as offensive
Urs Echternacht
Volunteer tester
Avatar

Send message
Joined: 18 Jan 06
Posts: 1038
Credit: 18,734,730
RAC: 0
Germany
Message 54167 - Posted: 8 May 2015, 22:34:02 UTC - in response to Message 54164.  

We have new Linux AP builds deployed, please test.

Looking through the result list of this host i can resume that the "bool2"-issue on NV-GPUs with newest drivers (newer than 347.xx) is solved.

Then we can proceed to release on main for these builds - right?

http://setiweb.ssl.berkeley.edu/beta/astropulse_v7_x86_64-pc-linux-gnu__opencl_nvidia_100.html
http://setiweb.ssl.berkeley.edu/beta/astropulse_v7_x86_64-pc-linux-gnu__opencl_ati_100.html
Look at the number of hosts that do run the 7.08 linux-apps currently. It might take some more time to get the necessary number of wus done for a release.

Eric, how many wus are needed for release to main ?
_\|/_
U r s
ID: 54167 · Report as offensive
Profile Raistmer
Volunteer tester
Avatar

Send message
Joined: 18 Aug 05
Posts: 2423
Credit: 15,878,738
RAC: 0
Russia
Message 54168 - Posted: 9 May 2015, 6:20:38 UTC - in response to Message 54167.  

We have new Linux AP builds deployed, please test.

Looking through the result list of this host i can resume that the "bool2"-issue on NV-GPUs with newest drivers (newer than 347.xx) is solved.

Then we can proceed to release on main for these builds - right?

http://setiweb.ssl.berkeley.edu/beta/astropulse_v7_x86_64-pc-linux-gnu__opencl_nvidia_100.html
http://setiweb.ssl.berkeley.edu/beta/astropulse_v7_x86_64-pc-linux-gnu__opencl_ati_100.html
Look at the number of hosts that do run the 7.08 linux-apps currently. It might take some more time to get the necessary number of wus done for a release.

Eric, how many wus are needed for release to main ?

Don't forget that last NV builds differ from prev ones very small. Only in that bool2 change. Hence I would not do full-scale testing again.
News about SETI opt app releases: https://twitter.com/Raistmer
ID: 54168 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 2 Jul 13
Posts: 505
Credit: 5,019,318
RAC: 10
United States
Message 54186 - Posted: 16 May 2015, 17:07:00 UTC - in response to Message 53868.  

Yes, after some battles with yet another database outage, we are back.

We've also got MacOS OpenCL SETI@home to test.

It seems we're back to receiving mostly trash for SETI@home Tasks. I have managed to log a few SETI@home tasks at beta besides those on Main. It has become apparent that on My Mac, and a number of others, the Stock Mac MB CPU App is Much slower than on other Platforms, close to twice as slow. Here are some recent comparisons,
Stock setiathome_7.00_i686-apple-darwin - 4 hours 40 min 49 sec
New MBv7_7.05r2899_sse3_x86_64-apple-darwin - 2 hours 19 min 44 sec
That is basically Twice as Fast.
If anyone wants to conduct their own comparisons the download is here, http://www.arkayn.us/forum/index.php?topic=130.msg4193#msg4193
Of course, you might have to test it on Main for now...
ID: 54186 · Report as offensive
Profile Raistmer
Volunteer tester
Avatar

Send message
Joined: 18 Aug 05
Posts: 2423
Credit: 15,878,738
RAC: 0
Russia
Message 54187 - Posted: 16 May 2015, 20:04:59 UTC

Worth to provide that build as new stock one then.
News about SETI opt app releases: https://twitter.com/Raistmer
ID: 54187 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 2 Jul 13
Posts: 505
Credit: 5,019,318
RAC: 10
United States
Message 54188 - Posted: 16 May 2015, 20:26:55 UTC - in response to Message 54187.  

Worth to provide that build as new stock one then.

I think it's worth looking into. Right now there is a problem with the AVX feature. The current App was compiled with SSE3, SSSE3, and SSE41 flags, when I add the AVX flag the App will not work with my Core2 CPU. As far as I can tell, you have to have a separate App to use the AVX feature. It would be nice to see the Apps tested on other machines...
ID: 54188 · Report as offensive
Profile Raistmer
Volunteer tester
Avatar

Send message
Joined: 18 Aug 05
Posts: 2423
Credit: 15,878,738
RAC: 0
Russia
Message 54189 - Posted: 17 May 2015, 7:54:44 UTC - in response to Message 54188.  

Worth to provide that build as new stock one then.

I think it's worth looking into. Right now there is a problem with the AVX feature. The current App was compiled with SSE3, SSSE3, and SSE41 flags, when I add the AVX flag the App will not work with my Core2 CPU. As far as I can tell, you have to have a separate App to use the AVX feature. It would be nice to see the Apps tested on other machines...

Core2 definitely can't support AVX. Moreover, AVX requires proper support from OS too, cause new registers are added to be saved at process context switch.
If you build from stock sources then there is testing for functions that could run on particular host. If you build from opt (AKv8) sources - then testing for aVX capabilities not implied, separate builds are needed.
To ensure testing on many different hosts binary should be sent to Eric as RC for beta project first. If it will work OK on beta hosts set then it could be made new stock indeed.
News about SETI opt app releases: https://twitter.com/Raistmer
ID: 54189 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 2 Jul 13
Posts: 505
Credit: 5,019,318
RAC: 10
United States
Message 54194 - Posted: 18 May 2015, 17:13:10 UTC - in response to Message 54189.  

...To ensure testing on many different hosts binary should be sent to Eric as RC for beta project first. If it will work OK on beta hosts set then it could be made new stock indeed.

Eric has a new PM...

Hmmmm, 27 Downloads and still No feedback. I guess that means it works. It works for me ;-)
ID: 54194 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 2 Jul 13
Posts: 505
Credit: 5,019,318
RAC: 10
United States
Message 54233 - Posted: 22 May 2015, 23:29:47 UTC

From this thread, Obtaining New GPU AP Apps for Linux
Anyone know what could be causing this error?
http://setiathome.berkeley.edu/result.php?resultid=4163376247
INFO: can't open binary kernel file: /usr/sbin/BOINC/projects/setiathome.berkeley.edu/AstroPulse_Kernels_r2751.cl_GeForceGTX750Ti.bin_V7_TWIN_FFA_34647, continue with recompile...
terminate called after throwing an instance of 'std::logic_error'
what(): basic_string::_S_construct null not valid
SIGABRT: abort called
Stack trace (18 frames):
../../projects/setiathome.berkeley.edu/astropulse_7.08_x86_64-pc-linux-gnu__opencl_nvidia_100(boinc_catch_signal+0x4d)[0x4c6fdd]
/lib64/libpthread.so.0(+0x10940)[0x7f75c2741940]
/lib64/libc.so.6(gsignal+0x37)[0x7f75c1768c37]
/lib64/libc.so.6(abort+0x16a)[0x7f75c176a34a]
/usr/lib64/libstdc++.so.6(_ZN9__gnu_cxx27__verbose_terminate_handlerEv+0x155)[0x7f75c1e5a5d5]
/usr/lib64/libstdc++.so.6(+0x5d776)[0x7f75c1e58776]
/usr/lib64/libstdc++.so.6(+0x5d7a3)[0x7f75c1e587a3]
/usr/lib64/libstdc++.so.6(+0x5d9c2)[0x7f75c1e589c2]
/usr/lib64/libstdc++.so.6(_ZSt19__throw_logic_errorPKc+0x67)[0x7f75c1ea9907]
/usr/lib64/libstdc++.so.6(_ZNSs12_S_constructIPKcEEPcT_S3_RKSaIcESt20forward_iterator_tag+0xa1)[0x7f75c1eb5011]
/usr/lib64/libstdc++.so.6(_ZNSsC1EPKcRKSaIcE+0x38)[0x7f75c1eb53a8]
../../projects/setiathome.berkeley.edu/astropulse_7.08_x86_64-pc-linux-gnu__opencl_nvidia_100[0x484a0c]
../../projects/setiathome.berkeley.edu/astropulse_7.08_x86_64-pc-linux-gnu__opencl_nvidia_100[0x48507f]
../../projects/setiathome.berkeley.edu/astropulse_7.08_x86_64-pc-linux-gnu__opencl_nvidia_100[0x47120f]
../../projects/setiathome.berkeley.edu/astropulse_7.08_x86_64-pc-linux-gnu__opencl_nvidia_100[0x461a14]
../../projects/setiathome.berkeley.edu/astropulse_7.08_x86_64-pc-linux-gnu__opencl_nvidia_100[0x46a205]
/lib64/libc.so.6(__libc_start_main+0xf5)[0x7f75c1754b75]
../../projects/setiathome.berkeley.edu/astropulse_7.08_x86_64-pc-linux-gnu__opencl_nvidia_100[0x40bd89]

Exiting...
ID: 54233 · Report as offensive
dsh1
Volunteer tester

Send message
Joined: 11 Dec 14
Posts: 18
Credit: 19,944
RAC: 0
United States
Message 54235 - Posted: 23 May 2015, 2:41:10 UTC - in response to Message 54233.  

From this thread, Obtaining New GPU AP Apps for Linux
Anyone know what could be causing this error?
http://setiathome.berkeley.edu/result.php?resultid=4163376247
INFO: can't open binary kernel file: /usr/sbin/BOINC/projects/setiathome.berkeley.edu/AstroPulse_Kernels_r2751.cl_GeForceGTX750Ti.bin_V7_TWIN_FFA_34647, continue with recompile...
terminate called after throwing an instance of 'std::logic_error'
what(): basic_string::_S_construct null not valid
SIGABRT: abort called
Stack trace (18 frames):
../../projects/setiathome.berkeley.edu/astropulse_7.08_x86_64-pc-linux-gnu__opencl_nvidia_100(boinc_catch_signal+0x4d)[0x4c6fdd]
/lib64/libpthread.so.0(+0x10940)[0x7f75c2741940]
/lib64/libc.so.6(gsignal+0x37)[0x7f75c1768c37]
/lib64/libc.so.6(abort+0x16a)[0x7f75c176a34a]
/usr/lib64/libstdc++.so.6(_ZN9__gnu_cxx27__verbose_terminate_handlerEv+0x155)[0x7f75c1e5a5d5]
/usr/lib64/libstdc++.so.6(+0x5d776)[0x7f75c1e58776]
/usr/lib64/libstdc++.so.6(+0x5d7a3)[0x7f75c1e587a3]
/usr/lib64/libstdc++.so.6(+0x5d9c2)[0x7f75c1e589c2]
/usr/lib64/libstdc++.so.6(_ZSt19__throw_logic_errorPKc+0x67)[0x7f75c1ea9907]
/usr/lib64/libstdc++.so.6(_ZNSs12_S_constructIPKcEEPcT_S3_RKSaIcESt20forward_iterator_tag+0xa1)[0x7f75c1eb5011]
/usr/lib64/libstdc++.so.6(_ZNSsC1EPKcRKSaIcE+0x38)[0x7f75c1eb53a8]
../../projects/setiathome.berkeley.edu/astropulse_7.08_x86_64-pc-linux-gnu__opencl_nvidia_100[0x484a0c]
../../projects/setiathome.berkeley.edu/astropulse_7.08_x86_64-pc-linux-gnu__opencl_nvidia_100[0x48507f]
../../projects/setiathome.berkeley.edu/astropulse_7.08_x86_64-pc-linux-gnu__opencl_nvidia_100[0x47120f]
../../projects/setiathome.berkeley.edu/astropulse_7.08_x86_64-pc-linux-gnu__opencl_nvidia_100[0x461a14]
../../projects/setiathome.berkeley.edu/astropulse_7.08_x86_64-pc-linux-gnu__opencl_nvidia_100[0x46a205]
/lib64/libc.so.6(__libc_start_main+0xf5)[0x7f75c1754b75]
../../projects/setiathome.berkeley.edu/astropulse_7.08_x86_64-pc-linux-gnu__opencl_nvidia_100[0x40bd89]

Exiting...



Before anyone tries to solve this, I did a lot poking around and found the "AstroPulse_Kernels_r2751.cl" file in /usr/sbin/BOINC/projects. After I copied it to /usr/sbin/BOINC/projects/setiathome.berkeley.edu and restarted boinc I found the "AstroPulse_Kernels_r2751.cl_GeForceGTX750Ti.bin_V7_TWIN_FFA_34647" had been created in /usr/sbin/BOINC/projects/setiathome.berkeley.edu.

The GPU is currently working on an AP WU with astropulse_7.08_x86_64-pc-linux-gnu__opencl_nvidia_100. The task is not finished yet so the jury is still out but the fact that it is still being worked on means it is a lot further along then previous attempts. I suspect everything is now fine so I would not waste any further time on this unless I report back that there is still a problem.
ID: 54235 · Report as offensive
Profile Raistmer
Volunteer tester
Avatar

Send message
Joined: 18 Aug 05
Posts: 2423
Credit: 15,878,738
RAC: 0
Russia
Message 54529 - Posted: 18 Jun 2015, 0:31:28 UTC

AstroPulse apps already released on main project. Currently MultiBeam apps under active testing.
Please switch from AstroPulse to MultiBeam processing on beta project.
News about SETI opt app releases: https://twitter.com/Raistmer
ID: 54529 · Report as offensive
TRuEQ & TuVaLu
Volunteer tester
Avatar

Send message
Joined: 28 Jan 11
Posts: 619
Credit: 2,580,051
RAC: 0
Sweden
Message 54830 - Posted: 10 Aug 2015, 20:10:12 UTC
Last modified: 10 Aug 2015, 20:36:41 UTC

ID: 54830 · Report as offensive
Previous · 1 . . . 32 · 33 · 34 · 35

Message boards : News : Astropulse 7,00 released for Linux 32&64, Win 32&64, Win32+AMD/NVIDIA/Intel GPU


 
©2020 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.