aboutsummaryrefslogtreecommitdiff
path: root/src/core/CL/cl_kernels/nchw/channel_shuffle.cl
diff options
context:
space:
mode:
authorPablo Marquez Tello <pablo.tello@arm.com>2021-10-18 10:50:31 +0100
committerPablo Marquez Tello <pablo.tello@arm.com>2021-10-18 13:11:42 +0000
commit36118524d2f387be53dc95e5eebabfcb3ec21f31 (patch)
tree4be226806f9dd74529b469f15b1254079b527c61 /src/core/CL/cl_kernels/nchw/channel_shuffle.cl
parent487d390b94ee93e1d89f5066ef8ca9442ab0a590 (diff)
downloadComputeLibrary-36118524d2f387be53dc95e5eebabfcb3ec21f31.tar.gz
Fix precision issue in ChannelShuffleKernel
* Fixed the issue in NHWC Neon * Fixed the rounding error in CL * Added a new test case to reproduce the problem * Resolves COMPMID-4831 Change-Id: I1613168cad580ca5acefe8ba340130af05cffaff Signed-off-by: Pablo Marquez Tello <pablo.tello@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/6454 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/core/CL/cl_kernels/nchw/channel_shuffle.cl')
-rw-r--r--src/core/CL/cl_kernels/nchw/channel_shuffle.cl4
1 files changed, 2 insertions, 2 deletions
diff --git a/src/core/CL/cl_kernels/nchw/channel_shuffle.cl b/src/core/CL/cl_kernels/nchw/channel_shuffle.cl
index 57d82e1e6f..84396e122f 100644
--- a/src/core/CL/cl_kernels/nchw/channel_shuffle.cl
+++ b/src/core/CL/cl_kernels/nchw/channel_shuffle.cl
@@ -33,7 +33,7 @@
#define DIV_MOD_UINT(x, y, div_res, mod_res) \
({ \
- div_res = (uint)((x) * (float)(1.0f / (float)(y))); \
+ div_res = (uint)((x)/(y)); \
uint r = div_res * (y); \
mod_res = (x)-r; \
})
@@ -100,4 +100,4 @@ __kernel void channel_shuffle_nchw(TENSOR4D_DECLARATION(src),
(u1, 0, (__global DATA_TYPE *)(output_ptr + 1 * dst_stride_y));
}
-#endif // defined(DATA_TYPE) && defined(VEC_SIZE) && defined(NUM_GROUPS) && defined(K) && defined(SRC_DIM_Z) \ No newline at end of file
+#endif // defined(DATA_TYPE) && defined(VEC_SIZE) && defined(NUM_GROUPS) && defined(K) && defined(SRC_DIM_Z)