diff options
author | Anthony Barbier <anthony.barbier@arm.com> | 2017-07-03 17:39:37 +0100 |
---|---|---|
committer | Anthony Barbier <anthony.barbier@arm.com> | 2018-09-17 14:15:39 +0100 |
commit | ac69aa137e360340fe9f148f019d93af6c3d8336 (patch) | |
tree | 88c2fb6ea8693c69b19d3b7b38fe26cd916303c8 /src/core/NEON/kernels/NELocallyConnectedMatrixMultiplyKernel.cpp | |
parent | 05da6dd102c988081c7d5eccb227f559f740ceef (diff) | |
download | ComputeLibrary-ac69aa137e360340fe9f148f019d93af6c3d8336.tar.gz |
COMPMID-418 Add check and fix comments after preprocessor conditions
Change-Id: I1353fd652ee180e3931e58b4ce13d651a48c7e2c
Reviewed-on: http://mpd-gerrit.cambridge.arm.com/79567
Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
Reviewed-by: Moritz Pflanzer <moritz.pflanzer@arm.com>
Diffstat (limited to 'src/core/NEON/kernels/NELocallyConnectedMatrixMultiplyKernel.cpp')
-rw-r--r-- | src/core/NEON/kernels/NELocallyConnectedMatrixMultiplyKernel.cpp | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/src/core/NEON/kernels/NELocallyConnectedMatrixMultiplyKernel.cpp b/src/core/NEON/kernels/NELocallyConnectedMatrixMultiplyKernel.cpp index ab84efbf23..895799c6ca 100644 --- a/src/core/NEON/kernels/NELocallyConnectedMatrixMultiplyKernel.cpp +++ b/src/core/NEON/kernels/NELocallyConnectedMatrixMultiplyKernel.cpp @@ -89,7 +89,7 @@ void vector_matrix_multiply_f32(const ITensor *input0, const ITensor *input1, IT asm volatile("PLD [%0, #128*4]" ::"r"(reinterpret_cast<const uint8_t *>(vec_a))); asm volatile("PLD [%0, #128*4]" ::"r"(reinterpret_cast<const uint8_t *>(matrix_b))); asm volatile("PLD [%0, #128*4]" ::"r"(reinterpret_cast<const uint8_t *>(matrix_b + in_b_stride))); -#endif +#endif /* __arm__ */ const float *vec_a_end_addr = vec_a + num_elems_vec_a; @@ -113,7 +113,7 @@ void vector_matrix_multiply_f32(const ITensor *input0, const ITensor *input1, IT asm volatile("PLD [%0, #128*1]" ::"r"(reinterpret_cast<const uint8_t *>(matrix_b + 2 * in_b_stride))); asm volatile("PLD [%0, #128*1]" ::"r"(reinterpret_cast<const uint8_t *>(matrix_b + 3 * in_b_stride))); asm volatile("PLD [%0, #128*1]" ::"r"(reinterpret_cast<const uint8_t *>(matrix_b + 4 * in_b_stride))); -#endif +#endif /* __arm __ */ acc0 = vmlaq_lane_f32(acc0, b00, a0l, 0); acc1 = vmlaq_lane_f32(acc1, b01, a0l, 0); |