aboutsummaryrefslogtreecommitdiff
path: root/src/core/CL/kernels/CLGEMMReshapeRHSMatrixKernel.cpp
diff options
context:
space:
mode:
authorGian Marco Iodice <gianmarco.iodice@arm.com>2019-01-11 11:30:55 +0000
committerGian Marco Iodice <gianmarco.iodice@arm.com>2019-01-15 15:06:12 +0000
commitbacfec5ecc3bd737c3d4eb2b0c165e0e55cc61f0 (patch)
tree12d115e3e158ac8bf434319fa9af44d75dc47785 /src/core/CL/kernels/CLGEMMReshapeRHSMatrixKernel.cpp
parent55e167814d462a803dbac82db17603cbe1258b4f (diff)
downloadComputeLibrary-bacfec5ecc3bd737c3d4eb2b0c165e0e55cc61f0.tar.gz
COMPMID-1687: Optimize CLGEMMMatrixMultiplyKernel (part 1)
Extended CLGEMMMatrixMultiplyReshapedKernel to support more parameters Change-Id: I4a27f986e3fe2dd071a4ccba5cfa0565f3db39ad Reviewed-on: https://review.mlplatform.org/495 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Diffstat (limited to 'src/core/CL/kernels/CLGEMMReshapeRHSMatrixKernel.cpp')
-rw-r--r--src/core/CL/kernels/CLGEMMReshapeRHSMatrixKernel.cpp7
1 files changed, 3 insertions, 4 deletions
diff --git a/src/core/CL/kernels/CLGEMMReshapeRHSMatrixKernel.cpp b/src/core/CL/kernels/CLGEMMReshapeRHSMatrixKernel.cpp
index 7cfa234b2e..7dc68201e5 100644
--- a/src/core/CL/kernels/CLGEMMReshapeRHSMatrixKernel.cpp
+++ b/src/core/CL/kernels/CLGEMMReshapeRHSMatrixKernel.cpp
@@ -1,5 +1,5 @@
/*
- * Copyright (c) 2018 ARM Limited.
+ * Copyright (c) 2018-2019 ARM Limited.
*
* SPDX-License-Identifier: MIT
*
@@ -47,12 +47,11 @@ Status validate_arguments(const ITensorInfo *input, const ITensorInfo *output, c
ARM_COMPUTE_RETURN_ERROR_ON(rhs_info.n0 == 0);
ARM_COMPUTE_RETURN_ERROR_ON(rhs_info.k0 == 0);
ARM_COMPUTE_RETURN_ERROR_ON(rhs_info.h0 == 0);
- ARM_COMPUTE_RETURN_ERROR_ON_MSG((rhs_info.n0 & (rhs_info.n0 - 1)), "Only power of two values are allowed for n0");
- ARM_COMPUTE_RETURN_ERROR_ON_MSG((rhs_info.k0 & (rhs_info.k0 - 1)), "Only power of two values are allowed for k0");
+ ARM_COMPUTE_RETURN_ERROR_ON_MSG(((rhs_info.n0 & (rhs_info.n0 - 1)) && rhs_info.n0 != 3), "Only 2,3,4,8,16 are supported for n0");
+ ARM_COMPUTE_RETURN_ERROR_ON_MSG(((rhs_info.k0 & (rhs_info.k0 - 1)) && (rhs_info.k0 != 1) && (rhs_info.k0 != 3)), "Only 1,2,3,4,8,16 are supported for k0");
ARM_COMPUTE_RETURN_ERROR_ON(rhs_info.n0 > 16);
ARM_COMPUTE_RETURN_ERROR_ON(rhs_info.k0 > 16);
ARM_COMPUTE_RETURN_ERROR_ON((rhs_info.k0 == 1) && (rhs_info.transpose));
- ARM_COMPUTE_RETURN_ERROR_ON((rhs_info.k0 == 2) && (rhs_info.transpose));
ARM_COMPUTE_RETURN_ERROR_ON_F16_UNSUPPORTED(input);
ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input, 1, DataType::QASYMM8, DataType::U8, DataType::S8,