diff options
| author | Gian Marco Iodice <gianmarco.iodice@arm.com> | 2019-06-04 12:39:23 +0100 |
|---|---|---|
| committer | Gian Marco Iodice <gianmarco.iodice@arm.com> | 2019-06-05 09:35:34 +0000 |
| commit | c625acd2a60a4fe34633c5cecef85c230933f772 (patch) | |
| tree | 22f55d82531be7ba889228f59f54aa901f36fbc0 /src/core/CL/kernels/CLGEMMMatrixMultiplyReshapedOnlyRHSKernel.cpp | |
| parent | b0f342ec315397e4b87d3a9cc3d12f3645c153bc (diff) | |
| download | ComputeLibrary-c625acd2a60a4fe34633c5cecef85c230933f772.tar.gz | |
COMPMID-2385: Fix CLGEMM
Change-Id: If796bc062a474ab941e48e1ae3819d5102084db4
Signed-off-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Reviewed-on: https://review.mlplatform.org/c/1280
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Giuseppe Rossini <giuseppe.rossini@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/core/CL/kernels/CLGEMMMatrixMultiplyReshapedOnlyRHSKernel.cpp')
| -rw-r--r-- | src/core/CL/kernels/CLGEMMMatrixMultiplyReshapedOnlyRHSKernel.cpp | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/src/core/CL/kernels/CLGEMMMatrixMultiplyReshapedOnlyRHSKernel.cpp b/src/core/CL/kernels/CLGEMMMatrixMultiplyReshapedOnlyRHSKernel.cpp index 58c4cdd2f6..f764b9d36d 100644 --- a/src/core/CL/kernels/CLGEMMMatrixMultiplyReshapedOnlyRHSKernel.cpp +++ b/src/core/CL/kernels/CLGEMMMatrixMultiplyReshapedOnlyRHSKernel.cpp @@ -154,7 +154,7 @@ std::pair<Status, Window> validate_and_configure_window(ITensorInfo *input0, ITe // Note: bottom paddings are calculated manually as the output can be reinterpreted as 3D tensor // The only way to set properly the paddings, it is to set those explicitly through the AccessWindowStatic - const int m = gemm_info.m(); + const int m = reinterpret_output_as_3d ? gemm_info.m() : input0->dimension(1); const int bottom_pad = (num_elems_processed_per_iteration_y - (m % num_elems_processed_per_iteration_y)) % num_elems_processed_per_iteration_y; win = calculate_max_window(tmp_info, Steps(num_elems_processed_per_iteration_x, num_elems_processed_per_iteration_y)); |