diff options
author | Radu Salavat <radu.salavat@arm.com> | 2024-04-12 12:26:50 +0000 |
---|---|---|
committer | Radu Salavat <radu.salavat@arm.com> | 2024-04-12 15:22:56 +0000 |
commit | cdce25bcb110a344e05364803f221f1645dc1f78 (patch) | |
tree | a4a2147d217b4dad68fa5b89624979245a6f88f2 /src | |
parent | cfca87b91def4f455630f2094447dc0500b6256c (diff) | |
download | ComputeLibrary-cdce25bcb110a344e05364803f221f1645dc1f78.tar.gz |
Accumulation in Cpu Gemm kernels is not supported for quantized kernels in aarch32. This patch guards the relevant tests.
Partially Resolves: ONCPUML-1442
Signed-off-by: Radu Salavat <radu.salavat@arm.com>
Change-Id: I8eed80db4b522185c3c50c13f0f701aa48961057
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/11410
Reviewed-by: Gunes Bayir <gunes.bayir@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src')
-rw-r--r-- | src/cpu/operators/CpuGemmLowpMatrixMultiplyCore.cpp | 3 |
1 files changed, 3 insertions, 0 deletions
diff --git a/src/cpu/operators/CpuGemmLowpMatrixMultiplyCore.cpp b/src/cpu/operators/CpuGemmLowpMatrixMultiplyCore.cpp index 94e86c6077..52d2f17dbf 100644 --- a/src/cpu/operators/CpuGemmLowpMatrixMultiplyCore.cpp +++ b/src/cpu/operators/CpuGemmLowpMatrixMultiplyCore.cpp @@ -347,6 +347,9 @@ Status CpuGemmLowpMatrixMultiplyCore::validate(const ITensorInfo *a, // When using accumulation(in place summation), for now, the only supported DataType for output is S32. if (gemm_info.accumulate()) { +#ifdef __arm__ + ARM_COMPUTE_RETURN_ERROR_MSG("Accumulation is not supported for armv7"); +#endif /* __arm__ */ ARM_COMPUTE_RETURN_ERROR_ON_MSG(gemm_info.gemmlowp_output_stage().type != GEMMLowpOutputStageType::NONE, "Accumulation is not supported for output QASYMM8/QASYMM8_SIGNED"); } |