[Csnd] Multicore - segmentation fault
Date | 2012-01-06 11:46 |
From | obaudouin |
Subject | [Csnd] Multicore - segmentation fault |
Hello! I try to use the -j flag with Csound5.15 (Fedora RT / compilation ok with buildMulticore=1) and the new parser. No matter with -j0 or -j1 (but not very useful) With -j2 and more it breaks: "Parsing successful! Inserting Parallelism Constructs into AST Csound tidy up: Erreur de segmentation" And, I wonder if the -j flag define a number of threads or set the number of cores to use? Many thanks and happy new year! Olivier. -- View this message in context: http://csound.1045644.n5.nabble.com/Multicore-segmentation-fault-tp5125317p5125317.html Sent from the Csound - General mailing list archive at Nabble.com. Send bugs reports to the Sourceforge bug tracker https://sourceforge.net/tracker/?group_id=81968&atid=564599 Discussions of bugs and features can be posted here To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound" |
Date | 2012-01-06 12:09 |
From | jpff@cs.bath.ac.uk |
Subject | Re: [Csnd] Multicore - segmentation fault |
> And, I wonder if the -j flag define a number of threads or set the number > of > cores to use? > > Threads Send bugs reports to the Sourceforge bug tracker https://sourceforge.net/tracker/?group_id=81968&atid=564599 Discussions of bugs and features can be posted here To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound" |
Date | 2012-01-06 12:15 |
From | obaudouin |
Subject | [Csnd] Re: Multicore - segmentation fault |
Thank you, but why the process breaks with -j > 1? -- View this message in context: http://csound.1045644.n5.nabble.com/Multicore-segmentation-fault-tp5125317p5125365.html Sent from the Csound - General mailing list archive at Nabble.com. Send bugs reports to the Sourceforge bug tracker https://sourceforge.net/tracker/?group_id=81968&atid=564599 Discussions of bugs and features can be posted here To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound" |
Date | 2012-01-06 12:25 |
From | jpff@cs.bath.ac.uk |
Subject | Re: [Csnd] Re: Multicore - segmentation fault |
> Thank you, but why the process breaks with -j > 1? > No idea. Which platform? What command-line options? Send bugs reports to the Sourceforge bug tracker https://sourceforge.net/tracker/?group_id=81968&atid=564599 Discussions of bugs and features can be posted here To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound" |
Date | 2012-01-06 12:32 |
From | obaudouin |
Subject | [Csnd] Re: Multicore - segmentation fault |
The options: |
Date | 2012-01-07 13:26 |
From | obaudouin |
Subject | [Csnd] Re: Multicore - segmentation fault |
Really no idea? ----- Olivier Baudouin, PhD MINT-OMF Paris-Sorbonne http://olivierbaudouin.com -- View this message in context: http://csound.1045644.n5.nabble.com/Multicore-segmentation-fault-tp5125317p5127696.html Sent from the Csound - General mailing list archive at Nabble.com. Send bugs reports to the Sourceforge bug tracker https://sourceforge.net/tracker/?group_id=81968&atid=564599 Discussions of bugs and features can be posted here To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound" |
Date | 2012-01-07 13:38 |
From | Victor |
Subject | Re: [Csnd] Re: Multicore - segmentation fault |
The only thing that springs to mind right now is the use of jack. Can you try output to file? Victor On 7 Jan 2012, at 13:26, obaudouin |
Date | 2012-01-07 13:48 |
From | obaudouin |
Subject | [Csnd] Re: Multicore - segmentation fault |
I have output to a file - without jack - but that doesn't solve the problem. ----- Olivier Baudouin, PhD MINT-OMF Paris-Sorbonne http://olivierbaudouin.com -- View this message in context: http://csound.1045644.n5.nabble.com/Multicore-segmentation-fault-tp5125317p5127718.html Sent from the Csound - General mailing list archive at Nabble.com. Send bugs reports to the Sourceforge bug tracker https://sourceforge.net/tracker/?group_id=81968&atid=564599 Discussions of bugs and features can be posted here To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound" |
Date | 2012-01-07 15:27 |
From | Tito Latini |
Subject | Re: [Csnd] Re: Multicore - segmentation fault |
Attachments | None |
Date | 2012-01-07 15:59 |
From | obaudouin |
Subject | [Csnd] Re: Multicore - segmentation fault |
This csd crashes when adding -j2 to the csound command line: |
Date | 2012-01-07 16:18 |
From | jpff@cs.bath.ac.uk |
Subject | Re: [Csnd] Re: Multicore - segmentation fault |
Could be thjat we did not expect an empty instrument so creating the DAG fails. Any chance of running with -v option and/or running under gdb ? > This csd crashes when adding -j2 to the csound command line: > > |
Date | 2012-01-07 16:30 |
From | obaudouin |
Subject | [Csnd] Re: Multicore - segmentation fault |
It crashes also with a non-empty instrument. WITH GDB: gdb csound (gdb) run -j2 test.csd (...) Leaving buffer 0x955390 Parsing successful! Inserting Parallelism Constructs into AST Program received signal SIGSEGV, Segmentation fault. 0x000000000041711b in ?? () ----- Olivier Baudouin, PhD MINT-OMF Paris-Sorbonne http://olivierbaudouin.com -- View this message in context: http://csound.1045644.n5.nabble.com/Multicore-segmentation-fault-tp5125317p5127908.html Sent from the Csound - General mailing list archive at Nabble.com. Send bugs reports to the Sourceforge bug tracker https://sourceforge.net/tracker/?group_id=81968&atid=564599 Discussions of bugs and features can be posted here To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound" |
Date | 2012-01-07 16:59 |
From | jpff@cs.bath.ac.uk |
Subject | Re: [Csnd] Re: Multicore - segmentation fault |
Hum; not sure how to tackle this. I am not seeing any problems here, OpenSuse 64bit. The gdb was not helpful -- do you have valgrind? That might give use some help All I can see from this is that it was inserting the locking code when it died ==John ff > It crashes also with a non-empty instrument. > > WITH GDB: > > gdb csound > (gdb) run -j2 test.csd > > (...) > > Leaving buffer 0x955390 > Parsing successful! > Inserting Parallelism Constructs into AST > > Program received signal SIGSEGV, Segmentation fault. > 0x000000000041711b in ?? () > > ----- > Olivier Baudouin, PhD > > MINT-OMF > Paris-Sorbonne > http://olivierbaudouin.com > -- > View this message in context: > http://csound.1045644.n5.nabble.com/Multicore-segmentation-fault-tp5125317p5127908.html > Sent from the Csound - General mailing list archive at Nabble.com. > > > Send bugs reports to the Sourceforge bug tracker > https://sourceforge.net/tracker/?group_id=81968&atid=564599 > Discussions of bugs and features can be posted here > To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe > csound" > > > > Send bugs reports to the Sourceforge bug tracker https://sourceforge.net/tracker/?group_id=81968&atid=564599 Discussions of bugs and features can be posted here To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound" |
Date | 2012-01-07 17:07 |
From | obaudouin |
Subject | [Csnd] Re: Multicore - segmentation fault |
With valgrind: valgrind csound -j2 test.csd (...) Leaving buffer 0x52591c0 Parsing successful! Inserting Parallelism Constructs into AST ==16581== Use of uninitialised value of size 8 ==16581== at 0x4170FD: ??? (in /usr/local/bin/csound) ==16581== by 0x42FBA7: ??? (in /usr/local/bin/csound) ==16581== by 0x410C9C: ??? (in /usr/local/bin/csound) ==16581== by 0x405CDF: ??? (in /usr/local/bin/csound) ==16581== by 0x31A402169C: (below main) (in /lib64/libc-2.14.90.so) ==16581== ==16581== Invalid read of size 8 ==16581== at 0x4170FD: ??? (in /usr/local/bin/csound) ==16581== by 0x42FBA7: ??? (in /usr/local/bin/csound) ==16581== by 0x410C9C: ??? (in /usr/local/bin/csound) ==16581== by 0x405CDF: ??? (in /usr/local/bin/csound) ==16581== by 0x31A402169C: (below main) (in /lib64/libc-2.14.90.so) ==16581== Address 0x8 is not stack'd, malloc'd or (recently) free'd ==16581== ==16581== Conditional jump or move depends on uninitialised value(s) ==16581== at 0x31A408B542: __GI___strcasecmp_l (in /lib64/libc-2.14.90.so) ==16581== by 0x31A4022532: __gconv_open (in /lib64/libc-2.14.90.so) ==16581== by 0x31A40304F9: _nl_find_msg (in /lib64/libc-2.14.90.so) ==16581== by 0x31A4030C94: __dcigettext (in /lib64/libc-2.14.90.so) ==16581== by 0x31A4069703: psignal (in /lib64/libc-2.14.90.so) ==16581== by 0x40A89A: ??? (in /usr/local/bin/csound) ==16581== by 0x31A40362FF: ??? (in /lib64/libc-2.14.90.so) ==16581== by 0x4170FC: ??? (in /usr/local/bin/csound) ==16581== by 0x42FBA7: ??? (in /usr/local/bin/csound) ==16581== by 0x410C9C: ??? (in /usr/local/bin/csound) ==16581== by 0x405CDF: ??? (in /usr/local/bin/csound) ==16581== by 0x31A402169C: (below main) (in /lib64/libc-2.14.90.so) ==16581== ==16581== Use of uninitialised value of size 8 ==16581== at 0x31A408CD54: __GI___strcasecmp_l (in /lib64/libc-2.14.90.so) ==16581== by 0x31A4022532: __gconv_open (in /lib64/libc-2.14.90.so) ==16581== by 0x31A40304F9: _nl_find_msg (in /lib64/libc-2.14.90.so) ==16581== by 0x31A4030C94: __dcigettext (in /lib64/libc-2.14.90.so) ==16581== by 0x31A4069703: psignal (in /lib64/libc-2.14.90.so) ==16581== by 0x40A89A: ??? (in /usr/local/bin/csound) ==16581== by 0x31A40362FF: ??? (in /lib64/libc-2.14.90.so) ==16581== by 0x4170FC: ??? (in /usr/local/bin/csound) ==16581== by 0x42FBA7: ??? (in /usr/local/bin/csound) ==16581== by 0x410C9C: ??? (in /usr/local/bin/csound) ==16581== by 0x405CDF: ??? (in /usr/local/bin/csound) ==16581== by 0x31A402169C: (below main) (in /lib64/libc-2.14.90.so) ==16581== ==16581== Use of uninitialised value of size 8 ==16581== at 0x31A408CD58: __GI___strcasecmp_l (in /lib64/libc-2.14.90.so) ==16581== by 0x31A4022532: __gconv_open (in /lib64/libc-2.14.90.so) ==16581== by 0x31A40304F9: _nl_find_msg (in /lib64/libc-2.14.90.so) ==16581== by 0x31A4030C94: __dcigettext (in /lib64/libc-2.14.90.so) ==16581== by 0x31A4069703: psignal (in /lib64/libc-2.14.90.so) ==16581== by 0x40A89A: ??? (in /usr/local/bin/csound) ==16581== by 0x31A40362FF: ??? (in /lib64/libc-2.14.90.so) ==16581== by 0x4170FC: ??? (in /usr/local/bin/csound) ==16581== by 0x42FBA7: ??? (in /usr/local/bin/csound) ==16581== by 0x410C9C: ??? (in /usr/local/bin/csound) ==16581== by 0x405CDF: ??? (in /usr/local/bin/csound) ==16581== by 0x31A402169C: (below main) (in /lib64/libc-2.14.90.so) ==16581== Csound tidy up: Erreur de segmentation ==16581== ==16581== HEAP SUMMARY: ==16581== in use at exit: 79,625 bytes in 181 blocks ==16581== total heap usage: 64,309 allocs, 64,128 frees, 12,038,118 bytes allocated ==16581== ==16581== LEAK SUMMARY: ==16581== definitely lost: 1,340 bytes in 32 blocks ==16581== indirectly lost: 35,433 bytes in 20 blocks ==16581== possibly lost: 0 bytes in 0 blocks ==16581== still reachable: 42,852 bytes in 129 blocks ==16581== suppressed: 0 bytes in 0 blocks ==16581== Rerun with --leak-check=full to see details of leaked memory ==16581== ==16581== For counts of detected and suppressed errors, rerun with: -v ==16581== Use --track-origins=yes to see where uninitialised values come from ==16581== ERROR SUMMARY: 5 errors from 5 contexts (suppressed: 2 from 2) ----- Olivier Baudouin, PhD MINT-OMF Paris-Sorbonne http://olivierbaudouin.com -- View this message in context: http://csound.1045644.n5.nabble.com/Multicore-segmentation-fault-tp5125317p5127974.html Sent from the Csound - General mailing list archive at Nabble.com. Send bugs reports to the Sourceforge bug tracker https://sourceforge.net/tracker/?group_id=81968&atid=564599 Discussions of bugs and features can be posted here To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound" |
Date | 2012-01-07 17:27 |
From | jpff@cs.bath.ac.uk |
Subject | Re: [Csnd] Re: Multicore - segmentation fault |
Thank you. I do not think we are much further forward. It is calling the vcode to locate global accesses but does not finish. Needless to say i am not seeing these errors or read errors etc. Do you have the ability to build from source? Without some debugging tools (compiled with -g) we are going to have difficulties in doing anything =John ff > With valgrind: > > valgrind csound -j2 test.csd > > (...) > > Leaving buffer 0x52591c0 > Parsing successful! > Inserting Parallelism Constructs into AST > ==16581== Use of uninitialised value of size 8 > ==16581== at 0x4170FD: ??? (in /usr/local/bin/csound) > ==16581== by 0x42FBA7: ??? (in /usr/local/bin/csound) > ==16581== by 0x410C9C: ??? (in /usr/local/bin/csound) > ==16581== by 0x405CDF: ??? (in /usr/local/bin/csound) > ==16581== by 0x31A402169C: (below main) (in /lib64/libc-2.14.90.so) > ==16581== > ==16581== Invalid read of size 8 > ==16581== at 0x4170FD: ??? (in /usr/local/bin/csound) > ==16581== by 0x42FBA7: ??? (in /usr/local/bin/csound) > ==16581== by 0x410C9C: ??? (in /usr/local/bin/csound) > ==16581== by 0x405CDF: ??? (in /usr/local/bin/csound) > ==16581== by 0x31A402169C: (below main) (in /lib64/libc-2.14.90.so) > ==16581== Address 0x8 is not stack'd, malloc'd or (recently) free'd > ==16581== > ==16581== Conditional jump or move depends on uninitialised value(s) > ==16581== at 0x31A408B542: __GI___strcasecmp_l (in > /lib64/libc-2.14.90.so) > ==16581== by 0x31A4022532: __gconv_open (in /lib64/libc-2.14.90.so) > ==16581== by 0x31A40304F9: _nl_find_msg (in /lib64/libc-2.14.90.so) > ==16581== by 0x31A4030C94: __dcigettext (in /lib64/libc-2.14.90.so) > ==16581== by 0x31A4069703: psignal (in /lib64/libc-2.14.90.so) > ==16581== by 0x40A89A: ??? (in /usr/local/bin/csound) > ==16581== by 0x31A40362FF: ??? (in /lib64/libc-2.14.90.so) > ==16581== by 0x4170FC: ??? (in /usr/local/bin/csound) > ==16581== by 0x42FBA7: ??? (in /usr/local/bin/csound) > ==16581== by 0x410C9C: ??? (in /usr/local/bin/csound) > ==16581== by 0x405CDF: ??? (in /usr/local/bin/csound) > ==16581== by 0x31A402169C: (below main) (in /lib64/libc-2.14.90.so) > ==16581== > ==16581== Use of uninitialised value of size 8 > ==16581== at 0x31A408CD54: __GI___strcasecmp_l (in > /lib64/libc-2.14.90.so) > ==16581== by 0x31A4022532: __gconv_open (in /lib64/libc-2.14.90.so) > ==16581== by 0x31A40304F9: _nl_find_msg (in /lib64/libc-2.14.90.so) > ==16581== by 0x31A4030C94: __dcigettext (in /lib64/libc-2.14.90.so) > ==16581== by 0x31A4069703: psignal (in /lib64/libc-2.14.90.so) > ==16581== by 0x40A89A: ??? (in /usr/local/bin/csound) > ==16581== by 0x31A40362FF: ??? (in /lib64/libc-2.14.90.so) > ==16581== by 0x4170FC: ??? (in /usr/local/bin/csound) > ==16581== by 0x42FBA7: ??? (in /usr/local/bin/csound) > ==16581== by 0x410C9C: ??? (in /usr/local/bin/csound) > ==16581== by 0x405CDF: ??? (in /usr/local/bin/csound) > ==16581== by 0x31A402169C: (below main) (in /lib64/libc-2.14.90.so) > ==16581== > ==16581== Use of uninitialised value of size 8 > ==16581== at 0x31A408CD58: __GI___strcasecmp_l (in > /lib64/libc-2.14.90.so) > ==16581== by 0x31A4022532: __gconv_open (in /lib64/libc-2.14.90.so) > ==16581== by 0x31A40304F9: _nl_find_msg (in /lib64/libc-2.14.90.so) > ==16581== by 0x31A4030C94: __dcigettext (in /lib64/libc-2.14.90.so) > ==16581== by 0x31A4069703: psignal (in /lib64/libc-2.14.90.so) > ==16581== by 0x40A89A: ??? (in /usr/local/bin/csound) > ==16581== by 0x31A40362FF: ??? (in /lib64/libc-2.14.90.so) > ==16581== by 0x4170FC: ??? (in /usr/local/bin/csound) > ==16581== by 0x42FBA7: ??? (in /usr/local/bin/csound) > ==16581== by 0x410C9C: ??? (in /usr/local/bin/csound) > ==16581== by 0x405CDF: ??? (in /usr/local/bin/csound) > ==16581== by 0x31A402169C: (below main) (in /lib64/libc-2.14.90.so) > ==16581== > Csound tidy up: Erreur de segmentation > ==16581== > ==16581== HEAP SUMMARY: > ==16581== in use at exit: 79,625 bytes in 181 blocks > ==16581== total heap usage: 64,309 allocs, 64,128 frees, 12,038,118 > bytes > allocated > ==16581== > ==16581== LEAK SUMMARY: > ==16581== definitely lost: 1,340 bytes in 32 blocks > ==16581== indirectly lost: 35,433 bytes in 20 blocks > ==16581== possibly lost: 0 bytes in 0 blocks > ==16581== still reachable: 42,852 bytes in 129 blocks > ==16581== suppressed: 0 bytes in 0 blocks > ==16581== Rerun with --leak-check=full to see details of leaked memory > ==16581== > ==16581== For counts of detected and suppressed errors, rerun with: -v > ==16581== Use --track-origins=yes to see where uninitialised values come > from > ==16581== ERROR SUMMARY: 5 errors from 5 contexts (suppressed: 2 from 2) > > ----- > > Olivier Baudouin, PhD > > MINT-OMF > Paris-Sorbonne > http://olivierbaudouin.com > -- Send bugs reports to the Sourceforge bug tracker https://sourceforge.net/tracker/?group_id=81968&atid=564599 Discussions of bugs and features can be posted here To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound" |
Date | 2012-01-07 17:36 |
From | Tito Latini |
Subject | Re: [Csnd] Re: Multicore - segmentation fault |
Attachments | None |
Date | 2012-01-07 17:55 |
From | Tito Latini |
Subject | Re: [Csnd] Re: Multicore - segmentation fault |
Attachments | None |
Date | 2012-01-07 18:01 |
From | obaudouin |
Subject | [Csnd] Re: Multicore - segmentation fault |
Yes, of course! It runs now. Thank you Tito. ----- Olivier Baudouin, PhD http://olivierbaudouin.com MINT-OMF Paris-Sorbonne -- View this message in context: http://csound.1045644.n5.nabble.com/Multicore-segmentation-fault-tp5125317p5128111.html Sent from the Csound - General mailing list archive at Nabble.com. Send bugs reports to the Sourceforge bug tracker https://sourceforge.net/tracker/?group_id=81968&atid=564599 Discussions of bugs and features can be posted here To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound" |
Date | 2012-01-07 18:11 |
From | jpff@cs.bath.ac.uk |
Subject | Re: [Csnd] Re: Multicore - segmentation fault |
Do you mean this is a named instrumnet problem? ==John ff > Yes, of course! It runs now. Thank you Tito. > > ----- > > Olivier Baudouin, PhD > http://olivierbaudouin.com > MINT-OMF > Paris-Sorbonne > > -- Send bugs reports to the Sourceforge bug tracker https://sourceforge.net/tracker/?group_id=81968&atid=564599 Discussions of bugs and features can be posted here To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound" |
Date | 2012-01-07 18:13 |
From | obaudouin |
Subject | [Csnd] Re: Multicore - segmentation fault |
yes. Olivier. Le 7 janv. 2012 19:12, "jpff [via Csound]" <[hidden email]> a écrit :
Do you mean this is a named instrumnet problem? Olivier Baudouin, PhD http://olivierbaudouin.com MINT-OMF Paris-Sorbonne View this message in context: Re: Multicore - segmentation fault Sent from the Csound - General mailing list archive at Nabble.com. |
Date | 2012-01-07 18:15 |
From | Tito Latini |
Subject | Re: [Csnd] Re: Multicore - segmentation fault |
Attachments | None |
Date | 2012-01-07 18:22 |
From | Tito Latini |
Subject | Re: [Csnd] Re: Multicore - segmentation fault |
Attachments | None |
Date | 2012-01-07 18:24 |
From | jpff@cs.bath.ac.uk |
Subject | Re: [Csnd] Re: Multicore - segmentation fault |
>> Do you mean this is a named instrumnet problem? > > yes, in cs_par_orc_semantic_analysis.c:348 > That is in get_by_num... getting more confused. Titom do you understand the problem? Send bugs reports to the Sourceforge bug tracker https://sourceforge.net/tracker/?group_id=81968&atid=564599 Discussions of bugs and features can be posted here To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound" |
Date | 2012-01-07 18:28 |
From | Tito Latini |
Subject | Re: [Csnd] Re: Multicore - segmentation fault |
Attachments | None |
Date | 2012-01-07 19:05 |
From | jpff@cs.bath.ac.uk |
Subject | Re: [Csnd] Re: Multicore - segmentation fault |
OK; perhaps we should move this to cs-dev > The problem is clear, the file, the line of reference > that I have sent you is wrong. Sorry, the last message > is correct > > tito > > On Sat, Jan 07, 2012 at 06:24:14PM -0000, jpff@cs.bath.ac.uk wrote: >> >> Do you mean this is a named instrumnet problem? >> > >> > yes, in cs_par_orc_semantic_analysis.c:348 >> > >> >> That is in get_by_num... getting more confused. Titom do you >> understand >> the problem? >> >> >> >> Send bugs reports to the Sourceforge bug tracker >> https://sourceforge.net/tracker/?group_id=81968&atid=564599 >> Discussions of bugs and features can be posted here >> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe >> csound" > > > Send bugs reports to the Sourceforge bug tracker > https://sourceforge.net/tracker/?group_id=81968&atid=564599 > Discussions of bugs and features can be posted here > To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe > csound" > > > > Send bugs reports to the Sourceforge bug tracker https://sourceforge.net/tracker/?group_id=81968&atid=564599 Discussions of bugs and features can be posted here To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound" |