已关闭
[Bug-Report|缺陷反馈]: ada_layer_norm_quant算子,执行单算子ut报错 #4207
liuhanwen8创建于  7月21日关闭于  11 天前
liuhanwen8
liuhanwen8
7月21日 创建

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ada_layer_norm_quant算子的ut,有ERROR

Environment / 环境信息 (Mandatory / 必填)

A3-9362
CANN 9.1.0-beta.3包
9.1.0-beta3分支代码

Steps to reproduce the issue / 重现步骤 (Mandatory / 必填)

执行bash build.sh -u --ops=ada_layer_norm_quant

Describe the expected behavior / 预期结果 (Mandatory / 必填)

无ERROR

[2026-07-21 16:35:04] [ RUN ] ada_layer_norm_quant_test.test_case_bfloat16_1
[2026-07-21 16:35:06] [TmSim]: Run in serial mode.
[2026-07-21 16:35:06] [ERROR][Block_2][AIV_0][ada_layer_norm_quant.cpp:109][CheckTensorAlignment][479243] Failed to check dst tensor start address alignment in ReduceRepeat, its start address must align with 8 bytes, current start address is 178180.
[2026-07-21 16:35:06]
[2026-07-21 16:35:06] [ERROR][CORE_2][pid 479243] error happened! =========
[2026-07-21 16:35:06] SIGABRT Signal (Abort Signal from abort) catched, backtrace info:
[2026-07-21 16:35:06] [#0] 0x00000000000bc8cf: AscendC::StackTrace() at ??:?
[2026-07-21 16:35:06] [#1] 0x00000000000bccb3: AscendC::BacktracePrint(int) at ??:?
[2026-07-21 16:35:06] [#2] 0x000000000044d583: Handler(int) at /home/l00797864/cann/91beta3/cann-9.1.0-beta.3/tools/cpudebug/include/kern_fwk.h:43
[2026-07-21 16:35:06] [#3] 0x00000000004988e3: void AscendC::CheckTensorAlignment(AscendC::LocalTensor const&, unsigned int, char const*, char const*) at /home/l00797864/cann/91beta3/cann-9.1.0-beta.3/aarch64-linux/asc/impl/basic_api/kernel_npu_debug.h:109 (discriminator 6)
[2026-07-21 16:35:06] [#4] 0x00000000004b3fa7: void AscendC::Internal::CheckReduceRepeatCommonParams<float, float>(AscendC::LocalTensor const&, AscendC::LocalTensor const&, int, AscendC::ReduceOrder, char const*) at /home/l00797864/cann/91beta3/cann-9.1.0-beta.3/aarch64-linux/asc/impl/basic_api/kernel_operator_vec_reduce_intf_impl.h:103
[2026-07-21 16:35:06] [#5] 0x00000000004ac7af: void AscendC::Internal::CheckReduceRepeatMaxMinParams<float, float, true, int>(AscendC::LocalTensor const&, AscendC::LocalTensor const&, int, int, int, int, int, AscendC::ReduceOrder) at /home/l00797864/cann/91beta3/cann-9.1.0-beta.3/aarch64-linux/asc/impl/basic_api/kernel_operator_vec_reduce_intf_impl.h:223
[2026-07-21 16:35:06] [#6] 0x00000000004a13fb: void AscendC::Internal::ReduceRepeatCommon<(AscendC::ReduceType)1, float, float, true, int>(AscendC::LocalTensor const&, AscendC::LocalTensor const&, int, int, int, int, int, AscendC::ReduceOrder) at /home/l00797864/cann/91beta3/cann-9.1.0-beta.3/aarch64-linux/asc/impl/basic_api/kernel_operator_vec_reduce_intf_impl.h:273
[2026-07-21 16:35:06] [#7] 0x00000000004954ff: void AscendC::ReduceRepeat<(AscendC::ReduceType)1, float, float, true>(AscendC::LocalTensor const&, AscendC::LocalTensor const&, int, int, int, int, int, AscendC::ReduceOrder) at /home/l00797864/cann/91beta3/cann-9.1.0-beta.3/aarch64-linux/asc/impl/basic_api/kernel_operator_vec_reduce_intf_impl.h:333
[2026-07-21 16:35:06] [#8] 0x0000000000488023: void AscendC::WholeReduceMax<float, true>(AscendC::LocalTensor const&, AscendC::LocalTensor const&, int, int, int, int, int, AscendC::ReduceOrder) at /home/l00797864/cann/91beta3/cann-9.1.0-beta.3/aarch64-linux/asc/impl/basic_api/kernel_operator_vec_reduce_intf_impl.h:616
[2026-07-21 16:35:06] [#9] 0x000000000047ee23: AdaLayerNormNS::ReduceMaxCustom(AscendC::LocalTensor const&, AscendC::LocalTensor const&, long) at /home/l00797864/cann/91beta3/ops-nn-9.1.0-beta.3/build/tbe/ascendc/ada_layer_norm_quant/../ada_layer_norm/ada_layer_norm_util.h:110 (discriminator 4)
[2026-07-21 16:35:06] [#10] 0x00000000004a95f3: AdaLayerNormNS::AdaLayerNormND<bfloat16::Bf16T, bfloat16::Bf16T, (unsigned char)2>::DynamicQuant(AdaLayerNormNS::RowRange, long) at /home/l00797864/cann/91beta3/ops-nn-9.1.0-beta.3/build/tbe/ascendc/ada_layer_norm_quant/../ada_layer_norm/ada_layer_norm_base_v1.h:297 (discriminator 3)
[2026-07-21 16:35:06] [#11] 0x000000000049ca0b: AdaLayerNormNS::AdaLayerNormND<bfloat16::Bf16T, bfloat16::Bf16T, (unsigned char)2>::FastProcess() at /home/l00797864/cann/91beta3/ops-nn-9.1.0-beta.3/build/tbe/ascendc/ada_layer_norm_quant/../ada_layer_norm/ada_layer_norm_base.h:337
[2026-07-21 16:35:06] [#12] 0x000000000048cfdf: AdaLayerNormNS::AdaLayerNormND<bfloat16::Bf16T, bfloat16::Bf16T, (unsigned char)2>::Process() at /home/l00797864/cann/91beta3/ops-nn-9.1.0-beta.3/build/tbe/ascendc/ada_layer_norm_quant/../ada_layer_norm/ada_layer_norm_base.h:187
[2026-07-21 16:35:06] [#13] 0x0000000000472db7: ada_layer_norm_quant at /home/l00797864/cann/91beta3/ops-nn-9.1.0-beta.3/build/tbe/ascendc/ada_layer_norm_quant/ada_layer_norm_quant.cpp:46 (discriminator 1)
[2026-07-21 16:35:06] [#14] 0x0000000000452447: void AscendC::RunKernelFunctionOnCpu<void ()(unsigned char, unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*), unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*>(void ()(unsigned char, unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*), char const*, unsigned int, unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*, unsigned char*) at /home/l00797864/cann/91beta3/cann-9.1.0-beta.3/tools/cpudebug/include/kern_fwk.h:142
[2026-07-21 16:35:06] [#15] 0x0000000000444b63: ada_layer_norm_quant_test_test_case_bfloat16_1_Test::TestBody() at /home/l00797864/cann/91beta3/ops-nn-9.1.0-beta.3/norm/ada_layer_norm_quant/tests/ut/op_kernel/test_ada_layer_norm_quant.cpp:81
[2026-07-21 16:35:06] [#16] 0x00000000005049f7: void testing::internal::HandleExceptionsInMethodIfSupported<testing::Test, void>(testing::Test*, void (testing::Test::)(), char const) at ??:?
[2026-07-21 16:35:06] [#17] 0x00000000004f3403: testing::Test::Run() at ??:?
[2026-07-21 16:35:06] [#18] 0x00000000004f360f: testing::TestInfo::Run() at ??:?
[2026-07-21 16:35:06] [#19] 0x00000000004f62f7: testing::TestSuite::Run() at ??:?
[2026-07-21 16:35:06] [#20] 0x00000000004f9aa3: testing::internal::UnitTestImpl::RunAllTests() at ??:?
[2026-07-21 16:35:06] [#21] 0x00000000004f37a3: testing::UnitTest::Run() at ??:?
[2026-07-21 16:35:06] [#22] 0x0000000000437087: RUN_ALL_TESTS() at /home/l00797864/cann/91beta3/ops-nn-9.1.0-beta.3/third_party/gtest/include/gtest/gtest.h:2317
[2026-07-21 16:35:06] [#23] 0x000000000042af9b: main at /home/l00797864/cann/91beta3/ops-nn-9.1.0-beta.3/tests/ut/op_kernel/test_op_kernel_main.cpp:66
[2026-07-21 16:35:06] [#24] 0x000000000042acaf: _start at ??:?

Special notes for this issue/备注 (Optional / 选填)

likedislike
yuning_chen
yuning_chen成员
7月21日 评论:

您好,感谢反馈,问题已收到,当前 @Chen_HaoWen 正在跟踪处理。

likedislike
yuning_chenyuning_chen成员
7月21日 将 Chen_HaoWen 设为负责人
CChen_HaoWen成员
7月31日 将 chenhaijie1423 设为负责人
Chen_HaoWen成员
7月31日 评论:

@chenhaijie1423 麻烦跟踪处理一下

likedislike
陈海杰成员
7月31日 关联了pull request:修复AdaLayerNormQuant跑UT报错的问题
陈海杰成员
7月31日 关联了pull request:修复AdaLayerNormQuant跑UT报错的问题
陈海杰成员
8月3日 评论:

@liuhanwen8 修复PR已合入,请更新代码重新验证下

likedislike
CChen_HaoWen成员
8月3日 添加了label:wait-feedback
CANN-robot
CANN-robot成员
26 天前 评论:

您好,为了更准确地定位和解决问题,我们需要您提供上述提到的相关信息。如果 14 天内没有进一步更新,我们将暂时关闭此 Issue。感谢您的理解与配合。

likedislike
CANN-robot
CANN-robot成员
11 天前 评论:

您好,由于该 Issue 已有一段时间没有更新,我们先将其关闭。如果您后续有新的进展或诉求,欢迎随时重新开启此 Issue 或提交新的 Issue,我们会继续为您跟进。

likedislike
CANN-robotCANN-robot成员
11 天前 关闭了 issue
CANN-robotCANN-robot成员
11 天前 添加了label:resolved