diff options
author | Viet-Hoa Do <viet-hoa.do@arm.com> | 2024-01-16 16:23:24 +0000 |
---|---|---|
committer | Viet-Hoa Do <viet-hoa.do@arm.com> | 2024-01-18 11:32:09 +0000 |
commit | 6829e0201e886cfd311e39f1d88f7452894bdfe5 (patch) | |
tree | ac1b97df4b76d0d09652ef3dbdce2be3d222df3c /src/core/CL/cl_kernels/common/slice_ops.cl | |
parent | 8896cf7cb7df34c699e7453a0f0c683d1202ed15 (diff) | |
download | ComputeLibrary-6829e0201e886cfd311e39f1d88f7452894bdfe5.tar.gz |
Fix divide-by-zero compilation error
* CONVERT_TO_TENSOR4D_STRUCT_NO_STEP is implemented and used
in some CL kernels in the way that causes divide-by-zero issue.
- Since the steps are all zeros, the issue might have been
ignored by the compiler.
Resolves: COMPMID-6795
Signed-off-by: Viet-Hoa Do <viet-hoa.do@arm.com>
Change-Id: I0fb38fc62d63671b8abefa39b3d9b3ca6f49c7fe
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/10967
Reviewed-by: Gunes Bayir <gunes.bayir@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/core/CL/cl_kernels/common/slice_ops.cl')
-rw-r--r-- | src/core/CL/cl_kernels/common/slice_ops.cl | 6 |
1 files changed, 3 insertions, 3 deletions
diff --git a/src/core/CL/cl_kernels/common/slice_ops.cl b/src/core/CL/cl_kernels/common/slice_ops.cl index d12c60f5ea..189d414aba 100644 --- a/src/core/CL/cl_kernels/common/slice_ops.cl +++ b/src/core/CL/cl_kernels/common/slice_ops.cl @@ -1,5 +1,5 @@ /* - * Copyright (c) 2018-2021 Arm Limited. + * Copyright (c) 2018-2021, 2024 Arm Limited. * * SPDX-License-Identifier: MIT * @@ -28,7 +28,7 @@ * @attention Supported tensor rank: up to 4 * * @attention Data type can be passed using the -DDATA_TYPE compile flag, e.g. -DDATA_TYPE=float - * @attention Input and output tensor dephts should be given as a preprocessor arguments using -DSRC_DEPTH=size. and -DDST_DEPTH=size + * @attention Output tensor depht should be given as a preprocessor arguments using -DDST_DEPTH=size * @attention Absolute start coordinates for each dimension should be given as preprocessor -DSTART_index=value e.g. -DSTART_0=2 * @attention Strides for each dimension should be given as preprocessor -DSTRIDE_index=value e.g. -DSTRIDE_1=1 * @@ -58,7 +58,7 @@ __kernel void strided_slice( TENSOR4D_DECLARATION(output)) { // Get pixels pointer - Tensor4D input = CONVERT_TO_TENSOR4D_STRUCT_NO_STEP(input, SRC_DEPTH); + Tensor4D input = CONVERT_TO_TENSOR4D_STRUCT_NO_STEP(input); Tensor4D output = CONVERT_TO_TENSOR4D_STRUCT(output, DST_DEPTH); int offset = 0; |