[Mesa-users] checking OSmesa performance

Daniel Fuster dfuster at gmail.com
Wed Jun 16 00:02:25 PDT 2010


Hello,

I am trying to use the Off-Screen Mesa (OSMesa) library in one cluster
to generate some graphics.

I have installed the version 7.8.1 using
./configure --prefix=$HOME --with-driver=osmesa
gmake

The installation seemed to work. Indeed,
$ pkg-config --modversion osmesa
7.8.1

I have also exported some env variables in my .bashrc
export LD_LIBRARY_PATH=$HOME/lib:$LD_LIBRARY_PATH
export GL_DIR=$HOME/lib
export OSMESA_DIR=${GL_DIR}

Then I have installed the software which is suposed to use the mesa
libraries. This program works perfectly in batch mode in other
machines and it seems to compile without problems

However, in  this case, when I try to use it I get a segmentation
fault (the output of valgrind is given below)

Could you tell me a way of checking if I have correctly installed the
OSmesa libraries without using my code? I cannot find any demo related
to that.

Can anybody tell me any clue about the error I am getting from the
debugger? As I say, this code has been working in other machines with
OSmesa libraries.

thanks for your help

Daniel

==392== Memcheck, a memory error detector
==392== Copyright (C) 2002-2009, and GNU GPL'd, by Julian Seward et al.
==392== Using Valgrind-3.5.0 and LibVEX; rerun with -h for copyright info
==392== Command: gfsview-batch2D file.gfs
==392==
==392== Invalid read of size 8
==392==    at 0x308C2512A7: glDisable (in /usr/lib64/libGL.so.1.2)
==392==    by 0x41B8B7: gfs_gl_init_gl (in
/autohome/u124/fusterd/bin/gfsview-batch2D)
==392==    by 0x40655B: read_commands (in
/autohome/u124/fusterd/bin/gfsview-batch2D)
==392==    by 0x406CC7: main (in /autohome/u124/fusterd/bin/gfsview-batch2D)
==392==  Address 0x6b99f90 is 64 bytes inside a block of size 72 free'd
==392==    at 0x4A06A4B: free (vg_replace_malloc.c:325)
==392==    by 0x5F71EF2: MPID_Dataloop_free (in
/apps/rhel5/mpich2-1.1.1p1/64/nemesis-gcc-4.4.0/lib/libmpich.so.1.1)
==392==    by 0x5F707B5: MPID_Dataloop_create_struct (in
/apps/rhel5/mpich2-1.1.1p1/64/nemesis-gcc-4.4.0/lib/libmpich.so.1.1)
==392==    by 0x5F70004: MPID_Dataloop_create_pairtype (in
/apps/rhel5/mpich2-1.1.1p1/64/nemesis-gcc-4.4.0/lib/libmpich.so.1.1)
==392==    by 0x5F95671: MPID_Type_create_pairtype (in
/apps/rhel5/mpich2-1.1.1p1/64/nemesis-gcc-4.4.0/lib/libmpich.so.1.1)
==392==    by 0x5FC8795: MPIR_Datatype_init (in
/apps/rhel5/mpich2-1.1.1p1/64/nemesis-gcc-4.4.0/lib/libmpich.so.1.1)
==392==    by 0x5F84934: MPIR_Init_thread (in
/apps/rhel5/mpich2-1.1.1p1/64/nemesis-gcc-4.4.0/lib/libmpich.so.1.1)
==392==    by 0x5F84445: PMPI_Init (in
/apps/rhel5/mpich2-1.1.1p1/64/nemesis-gcc-4.4.0/lib/libmpich.so.1.1)
==392==    by 0x4ED8DF6: gfs_init (init.c:357)
==392==    by 0x406A3C: main (in /autohome/u124/fusterd/bin/gfsview-batch2D)
==392==
==392== Jump to the invalid address stated on the next line
==392==    at 0x4C00080B: ???
==392==    by 0x41B8B7: gfs_gl_init_gl (in
/autohome/u124/fusterd/bin/gfsview-batch2D)
==392==    by 0x40655B: read_commands (in
/autohome/u124/fusterd/bin/gfsview-batch2D)
==392==    by 0x406CC7: main (in /autohome/u124/fusterd/bin/gfsview-batch2D)
==392==  Address 0x4c00080b is not stack'd, malloc'd or (recently) free'd
==392==
==392==
==392== Process terminating with default action of signal 11 (SIGSEGV)
==392==  Access not within mapped region at address 0x4C00080B
==392==    at 0x4C00080B: ???
==392==    by 0x41B8B7: gfs_gl_init_gl (in
/autohome/u124/fusterd/bin/gfsview-batch2D)
==392==    by 0x40655B: read_commands (in
/autohome/u124/fusterd/bin/gfsview-batch2D)
==392==    by 0x406CC7: main (in /autohome/u124/fusterd/bin/gfsview-batch2D)
==392==  If you believe this happened as a result of a stack
==392==  overflow in your program's main thread (unlikely but
==392==  possible), you can try to increase the size of the
==392==  main thread stack using the --main-stacksize= flag.
==392==  The main thread stack size used in this run was 16777216.
==392==
==392== HEAP SUMMARY:
==392==     in use at exit: 17,960,472 bytes in 3,506 blocks
==392==   total heap usage: 11,985 allocs, 8,479 frees, 19,238,812
bytes allocated
--


More information about the mesa-users mailing list