commit 8422558b946dc783f21241ed1226e601d314bd8b
author Georgios Pinitas <georgios.pinitas@arm.com> Mon May 14 12:00:05 2018 +0100
committer Anthony Barbier <anthony.barbier@arm.com> Fri Nov 02 16:52:19 2018 +0000
tree bd79b5a987dfb58b3dc4bf5297c25a6b1f346148
parent 393fa4c87c84356132303170d1b9ce9a45b3c3bf

COMPMID-1150 : (OCLGrind) Kernel compilation error and assertion

-Multiple definitions of COLS_MTX_B in gemm.cl one for FP32 and one for
FP16.
-GEMMTranspose1xWKernel invalid check fro small window sizes.

Change-Id: I9c7ddd3577aec9afc702731ca27a1e10d6eddb81
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/131023
Tested-by: Jenkins <bsgcomp@arm.com>
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>

src/core/CL/cl_kernels/gemm.cl

diff --git a/src/core/CL/cl_kernels/gemm.cl b/src/core/CL/cl_kernels/gemm.cl
index 7215f58..69bc09f 100644
--- a/src/core/CL/cl_kernels/gemm.cl
+++ b/src/core/CL/cl_kernels/gemm.cl
@@ -543,6 +543,9 @@
     vstore4((float4)(c30, c31, c32, c33), 0, (__global float *)(dst_addr + 3 * dst_stride_y));
 }
 
+// Undefine local defines
+#undef COLS_MTX_B
+
 #if defined(ARM_COMPUTE_OPENCL_FP16_ENABLED)
 /** This OpenCL kernel computes the matrix multiplication between matrix A (src0) and matrix B (src1)
  *  Matrix A and matrix B must be reshaped respectively with @ref gemm_interleave4x4_16bit and @ref gemm_transpose1x8 before running the matrix multiplication
@@ -879,6 +882,10 @@
     vstore8(c20, 0, (__global half *)(dst_addr + 2 * dst_stride_y));
     vstore8(c30, 0, (__global half *)(dst_addr + 3 * dst_stride_y));
 }
+
+// Undefine local defines
+#undef COLS_MTX_B
+
 #endif // defined(ARM_COMPUTE_OPENCL_FP16_ENABLED)
 
 #if defined(FIXED_POINT_POSITION)

src/core/CL/kernels/CLGEMMTranspose1xWKernel.cpp

diff --git a/src/core/CL/kernels/CLGEMMTranspose1xWKernel.cpp b/src/core/CL/kernels/CLGEMMTranspose1xWKernel.cpp
index a961860..05a20fd 100644
--- a/src/core/CL/kernels/CLGEMMTranspose1xWKernel.cpp
+++ b/src/core/CL/kernels/CLGEMMTranspose1xWKernel.cpp
@@ -71,11 +71,6 @@
     // Configure kernel window
     Window win = calculate_max_window(*input, Steps(num_elems_processed_per_iteration));
 
-    if((win.x().end() / scale_x) == 0)
-    {
-        return std::make_pair(ARM_COMPUTE_CREATE_ERROR(ErrorCode::RUNTIME_ERROR, "Transposed shape would be 0 in the second dimension"), win);
-    }
-
     AccessWindowHorizontal input_access(input, 0, num_elems_processed_per_iteration);
 
     // Output tensor auto inizialitation if not yet initialized