error - exit code 195 (0xc3) on CMSearch

Everything about the project RNA World
Nachricht
Autor
Benutzeravatar
yoyo
Vereinsvorstand
Vereinsvorstand
Beiträge: 8170
Registriert: 17.12.2002 14:09
Wohnort: Berlin

Re: error - exit code 195 (0xc3) on CMSearch

#25 Ungelesener Beitrag von yoyo » 09.03.2010 13:41

It seems to run on cisf system, which has a lot of swap space and needs ~3.3G.
So I will somehow add dynamic memory requirements to the WUs in the work generator.
yoyo
HILF mit im Rechenkraft-WiKi, dies gibts zu tun.
Wiki - FAQ - Verein - Chat

Bild Bild

Matthias Lehmkuhl
Prozessor-Polier
Prozessor-Polier
Beiträge: 125
Registriert: 12.03.2008 20:42

Re: error - exit code 195 (0xc3) on CMSearch

#26 Ungelesener Beitrag von Matthias Lehmkuhl » 09.03.2010 13:57

The Result of my Test with
# cmsearchsse3 :: search a sequence database with an RNA CM
...
# INFERNAL 1.0.2 (October 2009)Fatal exception (source file cp9_dp.c, line 841):
malloc of size 376230696 failed

This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.

Appended I have two screen shots from the Process Explorer properties from the Process (I think last Status of the process)
Du hast keine ausreichende Berechtigung, um die Dateianhänge dieses Beitrags anzusehen.
Matthias

Bild - Bild

cisf
Mikrocruncher
Mikrocruncher
Beiträge: 20
Registriert: 07.03.2010 21:55

Re: error - exit code 195 (0xc3) on CMSearch

#27 Ungelesener Beitrag von cisf » 09.03.2010 14:46

i ran the same work, with the windows program, on linux (wine installed)

Code: Alles auswählen

cisf@elisa:~/Desktop/test RNA_Boinc$ ./cmsearch cmfile in
# Z:\home\cisf\Desktop\test RNA_Boinc\cmsearch :: search a sequence database with an RNA CM
# INFERNAL 1.0.2 (October 2009)
# Copyright (C) 2009 HHMI Janelia Farm Research Campus
# Freely distributed under the GNU General Public License (GPLv3)
# - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
# command:    Z:\home\cisf\Desktop\test RNA_Boinc\cmsearch cmfile in
# date:       Tue Mar  9 15:14:14 2010
# compiler: 4.4.0, -O2 -fomit-frame-pointer -malign-double -fstrict-aliasing -specs=cm_gccspecs -mfpmath=387
# num seqs:   1
# dbsize(Mb): 188.115346
#
# Pre-search info for CM 1: 6S
#
#                                  cutoffs            predictions     
#                            -------------------  --------------------
# rnd  mod  alg  cfg   beta     E value   bit sc     surv     run time
# ---  ---  ---  ---  -----  ----------  -------  -------  -----------
    1  hmm  fwd  loc      -   27871.922     5.57   0.0491  00:59:20.54
    2   cm  cyk  loc  1e-010     702.737     6.78   0.0012  05:28:26.00
    3   cm  ins  loc  1e-015       1.000    20.41  9.1e-007  00:16:04.19
# ---  ---  ---  ---  -----  ----------  -------  -------  -----------
  all    -    -    -      -           -        -        -  06:43:50.74
#

CM: 6S
Fatal exception (source file cp9_dp.c, line 841):
malloc of size 376230696 failed

___________________
  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND

16337 cisf      20   0 2580m  32m 1968 R   91  1.8   1:20.42 cmsearch
16337 cisf      20   0 2580m 147m 1968 R  100  8.4  14:32.38 cmsearch 
16337 cisf      20   0 2580m  98m 1968 R  100  5.6  17:20.80 cmsearch
16337 cisf      20   0 2580m 170m 1968 R  100  9.7  20:46.43 cmsearch 




how is the code compiled for the windows machines ?

edit: added the pid resource usage

Ananas
WU-Schieber
WU-Schieber
Beiträge: 1184
Registriert: 27.04.2008 18:37
Wohnort: Nordlichter Köln

Re: error - exit code 195 (0xc3) on CMSearch

#28 Ungelesener Beitrag von Ananas » 09.03.2010 15:58

cisf hat geschrieben:...

Code: Alles auswählen

...
# compiler: 4.4.0, -O2 -fomit-frame-pointer -malign-double -fstrict-aliasing -specs=cm_gccspecs -mfpmath=387
...

how is the code compiled for the windows machines ?...
I knew this would be useful some day :-)
vi BOINC/checkin_notes
:1,$s/bug/feature/g
:wq!

Erzaehlen sich Biologen eigentlich Klein-RNA-Witze?

cisf
Mikrocruncher
Mikrocruncher
Beiträge: 20
Registriert: 07.03.2010 21:55

Re: error - exit code 195 (0xc3) on CMSearch

#29 Ungelesener Beitrag von cisf » 09.03.2010 19:11

offtopic:

i've made the cmsearch from the source (i haven't changed anything, just ./configure and make ) and :

Code: Alles auswählen

#
# expected time    actual time
# -------------  -------------
    03:22:48.12    01:46:34.00
//
#
# CPU time: 5070.93u 11.65s 01:24:42.58 Elapsed: 01:46:42
but when i ran the test with the aplication from the rna boinc folder (see previous posts) i got :

Code: Alles auswählen

#
# expected time    actual time
# -------------  -------------
    04:07:13.17    02:16:34.00
//
#
# CPU time: 7589.03u 10.58s 02:06:39.61 Elapsed: 02:16:40

i expected to see some difference, but isn't it to much !?
irelevant perhaps but also i have noticed that memory usage max. was 1.2g (compared to 1.3g)

ps: feel free to delete or move post

edit: the memory size diference note

Ananas
WU-Schieber
WU-Schieber
Beiträge: 1184
Registriert: 27.04.2008 18:37
Wohnort: Nordlichter Köln

Re: error - exit code 195 (0xc3) on CMSearch

#30 Ungelesener Beitrag von Ananas » 09.03.2010 19:51

For whatever reason it has decided you need the x32 version without SSE support - probably a wine issue? The .zip file in your project folder should contain better ones. For your system under Linux it used the sse2 version I think.

With configure it sure has choosen better optimizations, it might even have used a CPU specific "-march" parameter
vi BOINC/checkin_notes
:1,$s/bug/feature/g
:wq!

Erzaehlen sich Biologen eigentlich Klein-RNA-Witze?

cisf
Mikrocruncher
Mikrocruncher
Beiträge: 20
Registriert: 07.03.2010 21:55

Re: error - exit code 195 (0xc3) on CMSearch

#31 Ungelesener Beitrag von cisf » 09.03.2010 20:01

the compared times are both for the linux code ! (one from the project, with sse2; and one compiled on this machine - the faster one)

the windows code (run in linux with wine) errored the same way as to the ones who tested it on windows machines.

Ananas
WU-Schieber
WU-Schieber
Beiträge: 1184
Registriert: 27.04.2008 18:37
Wohnort: Nordlichter Köln

Re: error - exit code 195 (0xc3) on CMSearch

#32 Ungelesener Beitrag von Ananas » 09.03.2010 20:05

That's a bit confusing. It checks /proc/cpuinfo for sse support and should unzip the best version it can find for that cpu.

The unzip output seems to get lost though, so we cannot see which version it unzipped.

edit : Basically that cpu specific unzip stuff does work ...

http://www.rnaworld.de/rnaworld/downloa ... 36_438_0_1

(in this case an SSE2 box)

Could you please post your "flags" line from /proc/cpuinfo ?
vi BOINC/checkin_notes
:1,$s/bug/feature/g
:wq!

Erzaehlen sich Biologen eigentlich Klein-RNA-Witze?

Matthias Lehmkuhl
Prozessor-Polier
Prozessor-Polier
Beiträge: 125
Registriert: 12.03.2008 20:42

Re: error - exit code 195 (0xc3) on CMSearch

#33 Ungelesener Beitrag von Matthias Lehmkuhl » 09.03.2010 20:52

found one important info for Windows 32bit and 64bit on microsoft

http://msdn.microsoft.com/en-us/library ... 85%29.aspx

Memory type: User-mode virtual address space for each 32-bit process

Limit in 32-bit Windows: 2 GB

Up to 3 GB with IMAGE_FILE_LARGE_ADDRESS_AWARE and 4GT (how to set???)

Limit in 64-bit Windows (and 32bit process/program)
2 GB with IMAGE_FILE_LARGE_ADDRESS_AWARE cleared (default)

4 GB with IMAGE_FILE_LARGE_ADDRESS_AWARE set

for Linux I've seen something similar, but not found today

edit:
IMAGE_FILE_LARGE_ADDRESS_AWARE is part of the .exe ->
Executable files that can access the 3-GB address space must have the IMAGE_FILE_LARGE_ADDRESS_AWARE bit set in their image headers. ...
Matthias

Bild - Bild

Ananas
WU-Schieber
WU-Schieber
Beiträge: 1184
Registriert: 27.04.2008 18:37
Wohnort: Nordlichter Köln

Re: error - exit code 195 (0xc3) on CMSearch

#34 Ungelesener Beitrag von Ananas » 09.03.2010 21:44

creating new binaries ... will take a moment as I have to make clobber everything all the time (caused by the sse version make-scripts)


p.s.: no idea yet how to set "4GT" but this awareness thing is an option to the linker

always need a few p.s.'es : http://msdn.microsoft.com/en-us/library ... 85%29.aspx so 4GT is an OS setting, not a linker issue

This means, that just making the binary 4GB aware doesn't help in x86 mode, as long as boot.ini isn't tuned.

3rd p.s.: the 64 bit cross compiler doesn't accept --large-address-aware, so it might be default for those binaries.
vi BOINC/checkin_notes
:1,$s/bug/feature/g
:wq!

Erzaehlen sich Biologen eigentlich Klein-RNA-Witze?

Ananas
WU-Schieber
WU-Schieber
Beiträge: 1184
Registriert: 27.04.2008 18:37
Wohnort: Nordlichter Köln

Re: error - exit code 195 (0xc3) on CMSearch

#35 Ungelesener Beitrag von Ananas » 09.03.2010 22:19

One more thing I noticed : the archives for Linux x64 have a separate subdirectory for sse2 and there is a non-sse version under the default directory.

The sse2 version should reside in "default" and the sse2 subdirectory can be ditched.

One more thing ... the linux version of unzip_cpufeat in the server download directory seems to be not the latest version, which could be the reason for cisf's performance issues.
vi BOINC/checkin_notes
:1,$s/bug/feature/g
:wq!

Erzaehlen sich Biologen eigentlich Klein-RNA-Witze?

cisf
Mikrocruncher
Mikrocruncher
Beiträge: 20
Registriert: 07.03.2010 21:55

Re: error - exit code 195 (0xc3) on CMSearch

#36 Ungelesener Beitrag von cisf » 09.03.2010 22:27

Ananas hat geschrieben:(in this case an SSE2 box)
i think u wanted to say sse3 ( ssse3 on cpu list ) - cmsearch sse3 got properly extracted it seems
Ananas hat geschrieben:Could you please post your "flags" line from /proc/cpuinfo ?
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt rdtscp lm 3dnowext 3dnow rep_good extd_apicid pni cx16 lahf_lm cmp_legacy svm extapic cr8_legacy

mayby as u said ./configure did some particulare optimizing ...

Zurück zu „RNA World Discussions (english)“