2025-03-05T07:44:29.997285+08:00 rtx4080 kernel: pt_autograd_0[91731]: segfault at 42 ip 000074e6cac92b00 sp 000074e65aefb7d8 error 6 in libcuda.so.570.8
6.15[292b00,74e6cab77000+cf9000] likely on CPU 12 (core 24, socket 0)
2025-03-05T07:44:29.997643+08:00 rtx4080 kernel: Code: 85 d2 74 70 8b 87 70 02 00 00 49 89 fc 85 c0 74 63 49 89 f5 31 db eb 22 66 90 83 fa 02 75 7b 0f b7
06 66 89 07 83 c3 01 41 3b <9c> 24 70 02 00 00 73 42 49 8b 94 24 78 02 00 00 89 d8 48 8d 0c 40
nvidia-bug-report.log.gz (343.4 KB)
I’m currently using CUDA 12.8 and driver version 570.86.15. When I run deep learning tasks, I encounter the following issues. Could you please let me know how to resolve them? I’ve also provided related debugging reports and would appreciate any guidance on improvements. Thank you!