SIGSEGV, Segmentation fault in __cudaRegisterLinkedBinary

Hi guys,
I encountered problem with Segmentation fault in __cudaRegisterLinkedBinary.
Platform: xavier AGX 32GB eMMC
OS: ubuntu 18.04, cuda 10.0
docker: ubuntu 16.04

I debug this issue, it happened in “.so” library linking, not run into main function.
I am not clear, what introduce this?
just noticed function “_cudaRegisterLinkedBinary_38_cuda_device_runtime_compute_75
_cpp1_ii_8b1a5d37”, xavier is cmpute_72? but it’s privided by cuda installed by jetpack.

Thread 1 “yolo_example” received signal SIGSEGV, Segmentation fault.
0x0000007fb7f2c6fc in __cudaRegisterLinkedBinary(__fatBinC_Wrapper_t const*, void ()(void**), void) ()
from …so
(gdb) bt
#0 0x0000007fb7f2c6fc in __cudaRegisterLinkedBinary(__fatBinC_Wrapper_t const*, void ()(void**), void) ()
…so
#1 0x0000007fb7f2c69c in __cudaRegisterLinkedBinary_38_cuda_device_runtime_compute_75_cpp1_ii_8b1a5d37 ()
from …so
#2 0x0000007fb7fdfb60 in call_init (l=, argc=argc@entry=1,
argv=argv@entry=0x7fffffeee8, env=env@entry=0x7fffffeef8) at dl-init.c:72
#3 0x0000007fb7fdfca8 in call_init (env=0x7fffffeef8, argv=0x7fffffeee8, argc=1,
l=) at dl-init.c:30
#4 _dl_init (main_map=0x7fb8000190, argc=1, argv=0x7fffffeee8, env=0x7fffffeef8) at dl-init.c:120
#5 0x0000007fb7fd2d44 in _dl_start_user () from /lib/ld-linux-aarch64.so.1
Backtrace stopped: previous frame identical to this frame (corrupt stack?)

Hi,

Do you mind to share more information with us?

Do you meet this error when running an app within the docker container?
If yes, could you share which docker image do you use?

If not, may I know if this error is from a customized app or our official sample.
Could you also check if the sm=72 compute capacity is added when compiling the app?

Thanks.

issue fixed with comment off follow lines:

“-gencode=arch=compute_30,code=sm_30”,

“-gencode=arch=compute_50,code=sm_50”,

“-gencode=arch=compute_52,code=sm_52”,

“-gencode=arch=compute_60,code=sm_60”,

“-gencode=arch=compute_61,code=sm_61”,

“-gencode=arch=compute_61,code=compute_61”,

 "-gencode=arch=compute_72,code=sm_72",
"-gencode=arch=compute_72,code=compute_72",

Can some explain the root cause for it? Thanks ahead!

Hi,

Xavier’s GPU capacity is sm=72.
So you will need to add the correct architecture to generate a Xavier runnable kernel.

You can find all the GPU capacity information here:
https://developer.nvidia.com/cuda-gpus#compute

Thanks.