Thanks for sending an issue! Please fill in the following template to help quickly solve your problem.
执行算子非连续用例出现aicore
ascend950
跑用例必现
用例执行正常,且精度通过
[2026/05/19 22:57:49] 22:57:46 [INFO]-run_single.py:333 there is no case in /home/w00575733/heru n/op/ tcase/aclnn case/ e/aclnnscatterNdupdate to run,please check it. [2026/05/19 22:57:49] [INFO]-runsingtepy619startexecuteacnncase: acLnn5catterNdupdate aclinnScatterNdupdate int64 int32 ND acpw 000206 ungA [2026/05/19 22:57:49] [INFO]-acinScatterNdupdate.py:0246- [INF]-run acLnR:py:8798 - torch.npu.set-compilemode(jit compile-False)= -Start to run aclnn online- [2026/05/19 2 22:57:49] [2026/05/19 22:57:49] [INFO]-run acLnn-y:0713 3 - step:--. [INFO]-run acty:871 - stat4runing case aclnnscatternupdate aclnscatterNdupdte_ int64 int32N_acp_00206 ->>>>> Start to run online training on Npu <<<<<- [2026/05/19 22:57:49] [2026/05/19 22:57:49] [INFO]-run-aclnn.py:0715 - PID:649209 [2026-05-19 22:58:25] [WARNING]49209]proflerpy:Proflwontbusingwampthiscanskwproferresults [2026-05-19 22:58:25] eafdtreVctrchpu/sr/corenpu/strem.pp5funtonerr:devce errty3errocods55 [ERRR5225P4V)ERfd [5922:58:26.18539132 0pParamMaker.cpp:453]operator(:/torchu/src/aten/ops/papi/CopyKernel0pApi.cpp:9 Nu function error: device error type 3, error code is 507035 [EROR] 2026-05-19-22:58:26 (PID:649209, DeVice:0, RankID:-1) ER00100 PTA call acl api faled [Error]: The vector core execution is abnormal. Rectifythe fault based on the error information inthe ascend log. [ID:649209]226-5-19-22:58:26.3.911AclNRuntimeEror(Ez9):acrtLanchKrnelthostArgsfaldretrn:535 utnthsoletthenthfltocsandloatthfatadoth TraceBack (most recent call last): Failed to check task send status, stream id=42, retCode=0x715005e.[FUNC:StreamLaunchKernelV2][FILE:aix starsV2.cc][LINE:514] rtsLaunchKernelWithostArgsexecutonfalled,reason-vectorcoreexcepttonFUNcFuncEroreason[FILE:errormessage_manage.cc][LINE:5] atSLaunchernethstArsfafled,untmeresut-57[FUNC:ReportcalErrorFLE:log_nercpPLINE:148] aclrtlancherelwithostArgs faled, return: 507035 Launch kernel failed. KernelLaunchfailed: /home/w0057573/Ascend/can-9.1.0/opp/built-in/op_impl/ai_core/tbe//kernel/ascend950/ops_math/slice_apt/lice_a32d7ed6302227463347a4bd5a08_high_performance. Keed launch failed for slice, errno:361001. Exceptionraised from operator()at/torchnpu/csrc/aten/s/opapi/CpyKernelopApi.c:(mostrecent callfirst): fram #0: c10::Err r(c10::SourceLocation, std:: cxx11::bas ic c_string) + 0x98 (0x7c0ea158b218 in /npu/miniconda3/envs/CI_py310_aclnn/lib/python3.10/s
/assign
Thanks for sending an issue! Please fill in the following template to help quickly solve your problem.
Describe the current behavior / 问题描述 (Mandatory / 必填)
执行算子非连续用例出现aicore
Environment / 环境信息 (Mandatory / 必填)
ascend950
Steps to reproduce the issue / 重现步骤 (Mandatory / 必填)
跑用例必现
Describe the expected behavior / 预期结果 (Mandatory / 必填)
用例执行正常,且精度通过
Related log / screenshot / 日志 / 截图 (Mandatory / 必填)
[2026/05/19) + 0x98 (0x7c0ea158b218 in /npu/miniconda3/envs/CI_py310_aclnn/lib/python3.10/s
22:57:49]
22:57:46
[INFO]-run_single.py:333
there
is no case
in /home/w00575733/heru
n/op/
tcase/aclnn case/
e/aclnnscatterNdupdate to run,please check it.
[2026/05/19
22:57:49]
[INFO]-runsingtepy619startexecuteacnncase:
acLnn5catterNdupdate aclinnScatterNdupdate int64 int32 ND acpw 000206
ungA
[2026/05/19 22:57:49]
[INFO]-acinScatterNdupdate.py:0246-
[INF]-run acLnR:py:8798 -
torch.npu.set-compilemode(jit compile-False)=
-Start to run aclnn online-
[2026/05/19 2
22:57:49]
[2026/05/19
22:57:49]
[INFO]-run acLnn-y:0713
3 - step:--.
[INFO]-run acty:871 - stat4runing case aclnnscatternupdate aclnscatterNdupdte_ int64 int32N_acp_00206
->>>>> Start to run online training on Npu <<<<<-
[2026/05/19
22:57:49]
[2026/05/19
22:57:49]
[INFO]-run-aclnn.py:0715 - PID:649209
[2026-05-19 22:58:25]
[WARNING]49209]proflerpy:Proflwontbusingwampthiscanskwproferresults
[2026-05-19 22:58:25]
eafdtreVctrchpu/sr/corenpu/strem.pp5funtonerr:devce errty3errocods55
[ERRR5225P4V)ERfd
[5922:58:26.18539132 0pParamMaker.cpp:453]operator(:/torchu/src/aten/ops/papi/CopyKernel0pApi.cpp:9 Nu function error: device error type 3, error code is 507035
[EROR] 2026-05-19-22:58:26 (PID:649209, DeVice:0, RankID:-1) ER00100 PTA call acl api faled
[Error]: The vector core execution is abnormal.
Rectifythe fault based on the error information inthe ascend log.
[ID:649209]226-5-19-22:58:26.3.911AclNRuntimeEror(Ez9):acrtLanchKrnelthostArgsfaldretrn:535
utnthsoletthenthfltocsandloatthfatadoth
TraceBack (most recent call last):
Failed to check task send status, stream id=42, retCode=0x715005e.[FUNC:StreamLaunchKernelV2][FILE:aix starsV2.cc][LINE:514]
rtsLaunchKernelWithostArgsexecutonfalled,reason-vectorcoreexcepttonFUNcFuncEroreason[FILE:errormessage_manage.cc][LINE:5]
atSLaunchernethstArsfafled,untmeresut-57[FUNC:ReportcalErrorFLE:log_nercpPLINE:148]
aclrtlancherelwithostArgs faled, return: 507035
Launch kernel failed.
KernelLaunchfailed: /home/w0057573/Ascend/can-9.1.0/opp/built-in/op_impl/ai_core/tbe//kernel/ascend950/ops_math/slice_apt/lice_a32d7ed6302227463347a4bd5a08_high_performance.
Keed
launch failed for slice, errno:361001.
Exceptionraised from operator()at/torchnpu/csrc/aten/s/opapi/CpyKernelopApi.c:(mostrecent callfirst):
fram
#0:
c10::Err
r(c10::SourceLocation,
std::
cxx11::bas ic
c_string
Special notes for this issue/备注 (Optional / 选填)