diff options
| author | Jonathan Deakin <jonathan.deakin@arm.com> | 2022-08-22 13:32:37 +0100 |
|---|---|---|
| committer | Jonathan Deakin <jonathan.deakin@arm.com> | 2022-08-24 10:15:05 +0000 |
| commit | a331e48ad8a4856837cf0afdd44de69af43581af (patch) | |
| tree | d9f008f45d2918bf4259329940fc51963a0f8c6c /src/cpu/kernels/CpuAddKernel.cpp | |
| parent | 8bab588043da9a5b85fb549769b26a8bf24eaf5f (diff) | |
| download | ComputeLibrary-a331e48ad8a4856837cf0afdd44de69af43581af.tar.gz | |
Fix add for tensors with non-matching strides
Previously, the add_as_1d_array kernels were used on tensors with
non-matching strides which caused the wrong elements to be added. The
fix is to check that the strides are equal when selecting the addition
kernel.
Change-Id: I914ca2b95e5b8ed1875ec5ebe129bdfe2845496b
Signed-off-by: Jonathan Deakin <jonathan.deakin@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/8120
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Gunes Bayir <gunes.bayir@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/cpu/kernels/CpuAddKernel.cpp')
| -rw-r--r-- | src/cpu/kernels/CpuAddKernel.cpp | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/src/cpu/kernels/CpuAddKernel.cpp b/src/cpu/kernels/CpuAddKernel.cpp index 85ae410a94..8fc82ec184 100644 --- a/src/cpu/kernels/CpuAddKernel.cpp +++ b/src/cpu/kernels/CpuAddKernel.cpp @@ -41,7 +41,7 @@ namespace kernels { bool can_interpret_inputs_as_1d_array(const ITensorInfo &src0, const ITensorInfo &src1) { - return !src0.has_padding() && !src1.has_padding() && src0.tensor_shape() == src1.tensor_shape(); + return !src0.has_padding() && !src1.has_padding() && src0.tensor_shape() == src1.tensor_shape() && src0.strides_in_bytes() == src1.strides_in_bytes(); } namespace |