仓名: ops-math 算子名: conversion/depth_to_space 用例名: DepthToSpaceTiling.test_nchw_dcr_float16_success_001 UT类型: op_host
commit a3439d3c9(MR !6332「transpose 支持 ascend960dt/ascend960pr 平台及 4bit(HIF4/FP4)类型」)在 TransposeOpTilingData 中新增一个 int64 字段 bigDimBufferNum(conversion/transpose/op_host/arch35/transpose_tiling_arch35.h:97);DepthToSpaceTilingData 以内嵌结构复用 TransposeOpTilingData(conversion/depth_to_space/op_host/arch35/depth_to_space_tiling_arch35.h:26),raw tiling data 因此多出一个字段,而 depth_to_space 用例的 expectTilingData 未同步,字段整体错位一位,tiling_case_executor.cpp:274 比对 tiling data 失败(实际比期望多一个值)。
[2026-10-26 21:34:52] tiling_case_executor.cpp:274: Failure [2026-10-26 21:34:52] Expected equality of these values: [2026-10-26 21:34:52] Which is: "4 0 0 0 0 0 0 1 32 0 253952 1 0 0 0 0 0 0 0 1 2 2 4 2 0 0 0 0 4 2 2 2 0 0 0 0 2 0 3 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -1 0 0 0 0 0 3 1 4 2 1 2 2 4 2 1 4 2 2 2 1 2 2 0 2 1 0 2 2 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 2 0 2 1 0 2 2 2 " [2026-10-26 21:34:52] Which is: "4 0 0 0 0 0 0 1 32 0 253952 1 0 0 0 0 0 0 0 2 2 4 2 0 0 0 0 4 2 2 2 0 0 0 0 2 0 3 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -1 0 0 0 0 0 3 1 4 2 1 2 2 4 2 1 4 2 2 2 1 2 2 0 2 1 0 2 2 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 2 0 2 1 0 2 2 2 "
bash build.sh -u --opapi --ophost --cov --soc=ascend950
缺陷描述
仓名: ops-math
算子名: conversion/depth_to_space
用例名: DepthToSpaceTiling.test_nchw_dcr_float16_success_001
UT类型: op_host
失败原因
commit a3439d3c9(MR !6332「transpose 支持 ascend960dt/ascend960pr 平台及 4bit(HIF4/FP4)类型」)在 TransposeOpTilingData 中新增一个 int64 字段 bigDimBufferNum(conversion/transpose/op_host/arch35/transpose_tiling_arch35.h:97);DepthToSpaceTilingData 以内嵌结构复用 TransposeOpTilingData(conversion/depth_to_space/op_host/arch35/depth_to_space_tiling_arch35.h:26),raw tiling data 因此多出一个字段,而 depth_to_space 用例的 expectTilingData 未同步,字段整体错位一位,tiling_case_executor.cpp:274 比对 tiling data 失败(实际比期望多一个值)。
触发PR(agent 分析确认)
排查依据
失败日志摘录
复现命令