1
运行时
Memory map
2022-06-05 19:19:36
52

上一次HIP还能成功提交作业运行,下一次用surn命令提交作业的时候出现:

srun: ROUTE: split_hostlist: hl=e10r4n03 tree_width 0
*** Error in `/public/home/ghfund_1_a5/barhip/./kernel': free(): invalid pointer: 0x0000000001d10dc8 ***
======= Backtrace: =========
/lib64/libc.so.6(+0x81489)[0x2b9681e49489]
/public/software/compiler/rocm/dtk-21.04/lib/libamdhip64.so.4(+0x1627b6)[0x2b9680e397b6]
/public/software/compiler/rocm/dtk-21.04/lib/libamdhip64.so.4(+0x1624ba)[0x2b9680e394ba]
/public/software/compiler/rocm/dtk-21.04/lib/libamdhip64.so.4(+0x162860)[0x2b9680e39860]
/public/software/compiler/rocm/dtk-21.04/lib/libamdhip64.so.4(+0x9681a)[0x2b9680d6d81a]
/public/software/compiler/rocm/dtk-21.04/lib/libamdhip64.so.4(hipMemcpy+0x466)[0x2b9680da02f6]
/public/home/ghfund_1_a5/barhip/./kernel[0x401c9a]
/lib64/libc.so.6(__libc_start_main+0xf5)[0x2b9681dea3d5]
/public/home/ghfund_1_a5/barhip/./kernel[0x400caa]
======= Memory map: ========
00400000-0040d000 r-xp 00000000 00:33 324282659100 /public/home/ghfund_1_a5/barhip/kernel
0060c000-0060d000 r--p 0000c000 00:33 324282659100 /public/home/ghfund_1_a5/barhip/kernel
0060d000-0060e000 rw-p 0000d000 00:33 324282659100 /public/home/ghfund_1_a5/barhip/kernel
01bcb000-01d23000 rw-p 00000000 00:00 0 [heap]

全部评论(1)
只看楼主
T
tianlh
1楼

dtk21.10.1也有这个问题吗?尽量用sbatch脚本提交吧,您试下dtk21.10.1,还有问题的话贴一下作业脚本截图吧

2022/6/6
没有更多内容了
--
楼主
Level.1
0 帖子
楼主热帖
暂无数据