commit ae99b6eac40c7c3cb5eb465f3cbe4b522eff0488
author Giorgio Arena <giorgio.arena@arm.com> Thu Aug 01 14:22:12 2019 +0100
committer Georgios Pinitas <georgios.pinitas@arm.com> Mon Sep 02 15:31:29 2019 +0000
tree 1cd14abfd10953686185d3697c545830f26ac7bb
parent c2a60593436387d20ff142a619f4c3955a5cd41b

COMPMID-1965 Extend CLGEMMMatrixMultiplyReshapedKernel to support transposed LHS (t) and not-transpose RHS

Change-Id: I437a00d7213fefd6f4365071b46174d44df8b85c
Signed-off-by: Giorgio Arena <giorgio.arena@arm.com>
Reviewed-on: https://review.mlplatform.org/c/1677
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>

src/core/CL/kernels/CLGEMMMatrixMultiplyReshapedKernel.cpp

diff --git a/src/core/CL/kernels/CLGEMMMatrixMultiplyReshapedKernel.cpp b/src/core/CL/kernels/CLGEMMMatrixMultiplyReshapedKernel.cpp
index 63451b4..9630cae 100644
--- a/src/core/CL/kernels/CLGEMMMatrixMultiplyReshapedKernel.cpp
+++ b/src/core/CL/kernels/CLGEMMMatrixMultiplyReshapedKernel.cpp
@@ -67,9 +67,8 @@
     ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input0, input1);
     ARM_COMPUTE_RETURN_ERROR_ON_MSG(input0->num_dimensions() > 4, "The number of dimensions for the LHS matrix must be <= 4");
     ARM_COMPUTE_RETURN_ERROR_ON_MSG(input1->num_dimensions() > 3, "The number of dimensions for the RHS matrix must be <= 3");
-    ARM_COMPUTE_RETURN_ERROR_ON(lhs_info.transpose);
-    ARM_COMPUTE_RETURN_ERROR_ON(!rhs_info.transpose);
     ARM_COMPUTE_RETURN_ERROR_ON(lhs_info.k0 != rhs_info.k0);
+    ARM_COMPUTE_RETURN_ERROR_ON(lhs_info.transpose == rhs_info.transpose);
     ARM_COMPUTE_RETURN_ERROR_ON_MSG(((lhs_info.k0 & (lhs_info.k0 - 1)) && lhs_info.k0 != 3), "Only 2,3,4,8,16 are supported for k0");
     ARM_COMPUTE_RETURN_ERROR_ON(lhs_info.k0 > 16);
     ARM_COMPUTE_RETURN_ERROR_ON(lhs_info.m0 < 2 || lhs_info.m0 > 8);
@@ -253,6 +252,7 @@
     build_opts.add_option_if(!_slide_matrix_b, "-DMATRIX_B_DEPTH=" + support::cpp11::to_string(input1->info()->dimension(2)));
     build_opts.add_option_if(lhs_info.interleave, "-DLHS_INTERLEAVE");
     build_opts.add_option_if(rhs_info.interleave, "-DRHS_INTERLEAVE");
+    build_opts.add_option_if(lhs_info.transpose, "-DLHS_TRANSPOSE");
     build_opts.add_option_if(_use_dummy_work_items, "-DDUMMY_WORK_ITEMS");
     build_opts.add_option("-DM=" + support::cpp11::to_string(gemm_info.m));
     build_opts.add_option("-DN=" + support::cpp11::to_string(gemm_info.n));