Hi FS experts,
I'm currently testing FS v5 with cuda support.
I'm using Centos 5.5 x86_64 on a DELL XPS with 4GB RAM, Intel Core 2 Extreme Processor QX9650 @ 3.67GHz and three GPUS Geforce 9800 GT.
Every GPU has 512 GDDR3 memory, 256-BIT Memory Bus, 112 Stream Processing Units, 2x 400MHz RAMDACs and PCI Express 2.0 suport.
The SDK examples runs perfect but when I run the following script:
export FREESURFER_HOME=/home/neuro/freesurfer source /home/neuro/freesurfer/SetUpFreeSurfer.sh export SUBJECTS_DIR=/home/neuro/proc/OutPBS_GPU recon-all -all -force -i /home/neuro/orig/ACR_103/ACR_103_0002 -s ACR_103 -qcache *-use-cuda*
I obtain the following error message:
*** glibc detected *** mri_em_register: malloc(): memory corruption: 0x00000000315dfdc0 *** ======= Backtrace: ========= /lib64/libc.so.6[0x3304072fae] /lib64/libc.so.6(__libc_calloc+0xcd)[0x330407495d] mri_em_register[0x42efad] mri_em_register[0x40e80d] /lib64/libc.so.6(__libc_start_main+0xf4)[0x330401d994] mri_em_register(__gxx_personality_v0+0x16a)[0x40b16a] ======= Memory map: ======== 00400000-00b55000 r-xp 00400000 00:00 0 00b55000-00c54000 ---p 00b55000 00:00 0 00c54000-02627000 rwxp 00c54000 00:00 0 1ef6f000-3c61e000 rwxp 1ef6f000 00:00 0 [heap] 3303c00000-3303c1c000 r-xp 00000000 08:03 79331333 /lib64/ld-2.5.so 3303c1c000-3303e1b000 ---p 3303c1c000 00:00 0 3303e1b000-3303e1c000 r-xp 0001b000 08:03 79331333 /lib64/ld-2.5.so 3303e1c000-3303e1d000 rwxp 0001c000 08:03 79331333 /lib64/ld-2.5.so 3304000000-330414e000 r-xp 00000000 08:03 79331340 /lib64/libc-2.5.so 330414e000-330434d000 ---p 0014e000 08:03 79331340 /lib64/libc-2.5.so 330434d000-3304351000 r-xp 0014d000 08:03 79331340 /lib64/libc-2.5.so 3304351000-3304352000 rwxp 00151000 08:03 79331340 /lib64/libc-2.5.so 3304352000-3304357000 rwxp 3304352000 00:00 0 3304400000-3304402000 r-xp 00000000 08:03 79331344 /lib64/libdl-2.5.so 3304402000-3304602000 ---p 00002000 08:03 79331344 /lib64/libdl-2.5.so 3304602000-3304603000 r-xp 00002000 08:03 79331344 /lib64/libdl-2.5.so 3304603000-3304604000 rwxp 00003000 08:03 79331344 /lib64/libdl-2.5.so 3304800000-3304882000 r-xp 00000000 08:03 79331372 /lib64/libm-2.5.so 3304882000-3304a81000 ---p 00082000 08:03 79331372 /lib64/libm-2.5.so 3304a81000-3304a82000 r-xp 00081000 08:03 79331372 /lib64/libm-2.5.so 3304a82000-3304a83000 rwxp 00082000 08:03 79331372 /lib64/libm-2.5.so 3304c00000-3304c16000 r-xp 00000000 08:03 79331354 /lib64/libpthread-2.5.so 3304c16000-3304e15000 ---p 00016000 08:03 79331354 /lib64/libpthread-2.5.so 3304e15000-3304e16000 r-xp 00015000 08:03 79331354 /lib64/libpthread-2.5.so 3304e16000-3304e17000 rwxp 00016000 08:03 79331354 /lib64/libpthread-2.5.so 3304e17000-3304e1b000 rwxp 3304e17000 00:00 0 3305000000-3305014000 r-xp 00000000 08:03 12817044 /usr/lib64/libz.so.1.2.3 3305014000-3305213000 ---p 00014000 08:03 12817044 /usr/lib64/libz.so.1.2.3 3305213000-3305214000 rwxp 00013000 08:03 12817044 /usr/lib64/libz.so.1.2.3 3306400000-3306409000 r-xp 00000000 08:03 79331382 /lib64/libcrypt-2.5.so 3306409000-3306608000 ---p 00009000 08:03 79331382 /lib64/libcrypt-2.5.so 3306608000-3306609000 r-xp 00008000 08:03 79331382 /lib64/libcrypt-2.5.so 3306609000-330660a000 rwxp 00009000 08:03 79331382 /lib64/libcrypt-2.5.so 330660a000-3306638000 rwxp 330660a000 00:00 0 3309c00000-3309c0d000 r-xp 00000000 08:03 79331394 /lib64/libgcc_s-4.1.2-20080825.so.1 3309c0d000-3309e0d000 ---p 0000d000 08:03 79331394 /lib64/libgcc_s-4.1.2-20080825.so.1 3309e0d000-3309e0e000 rwxp 0000d000 08:03 79331394 /lib64/libgcc_s-4.1.2-20080825.so.1 330b400000-330b4e6000 r-xp 00000000 08:03 12824478 /usr/lib64/libstdc++.so.6.0.8 330b4e6000-330b6e5000 ---p 000e6000 08:03 12824478 /usr/lib64/libstdc++.so.6.0.8 330b6e5000-330b6eb000 r-xp 000e5000 08:03 12824478 /usr/lib64/libstdc++.so.6.0.8 330b6eb000-330b6ee000 rwxp 000eb000 08:03 12824478 /usr/lib64/libstdc++.so.6.0.8 330b6ee000-330b700000 rwxp 330b6ee000 00:00 0 2b2d46aaa000-2b2d46aad000 rwxp 2b2d46aaa000 00:00 0 2b2d46ab9000-2b2d47f2e000 rwxp 2b2d46ab9000 00:00 0 2b2d48000000-2b2d48021000 rwxp 2b2d48000000 00:00 0 2b2d48021000-2b2d4c000000 ---p 2b2d48021000 00:00 0 7fff73ffe000-7fff741f2000 rwxp 7fffffe0a000 00:00 0 [stack] ffffffffff600000-ffffffffffe00000 ---p 00000000 00:00 0 [vdso]
I attached the output log of the execution.
Any idea will be apreciated!
Thank you in advance!!
On Wed, 2010-10-20 at 16:18 +0200, Jordi Delgado wrote:
*** glibc detected *** mri_em_register: malloc(): memory corruption: 0x00000000315dfdc0 *** ======= Backtrace: ========= /lib64/libc.so.6[0x3304072fae] /lib64/libc.so.6(__libc_calloc+0xcd)[0x330407495d] mri_em_register[0x42efad] mri_em_register[0x40e80d] /lib64/libc.so.6(__libc_start_main+0xf4)[0x330401d994] mri_em_register(__gxx_personality_v0+0x16a)[0x40b16a]
<snip>
That's odd. However, mri_em_register is the plain CPU binary - the CUDA binaries have _cuda appended. Do you get the same problem if you don't enable CUDA when running mri_em_register?
Regards,
Richard
freesurfer@nmr.mgh.harvard.edu