aboutsummaryrefslogtreecommitdiff
path: root/src/core/NEON/kernels/arm_conv/depthwise/interleaves/a64_s8q_3x3_dot.cpp
diff options
context:
space:
mode:
authorMichael Tyler <michael.tyler@arm.com>2023-06-30 11:26:05 +0100
committermichael.tyler <michael.tyler@arm.com>2023-07-04 14:34:58 +0000
commit8deee9bd9b9137c256c23b86be11dbf0466f3aa8 (patch)
treeac80b3bdd992552b65e306b77f061484da0591ca /src/core/NEON/kernels/arm_conv/depthwise/interleaves/a64_s8q_3x3_dot.cpp
parent19844f605f5e5b71d05164711dee13f8652adafe (diff)
downloadComputeLibrary-8deee9bd9b9137c256c23b86be11dbf0466f3aa8.tar.gz
Depthwise channel pre-multiplication
Resolves: COMPMID-6337 Change-Id: Ie9097b3f56e8071426c621386a5988bd7f7e8ef2 Signed-off-by: Michael Tyler <michael.tyler@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/9852 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Viet-Hoa Do <viet-hoa.do@arm.com> Benchmark: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/core/NEON/kernels/arm_conv/depthwise/interleaves/a64_s8q_3x3_dot.cpp')
-rw-r--r--src/core/NEON/kernels/arm_conv/depthwise/interleaves/a64_s8q_3x3_dot.cpp2
1 files changed, 1 insertions, 1 deletions
diff --git a/src/core/NEON/kernels/arm_conv/depthwise/interleaves/a64_s8q_3x3_dot.cpp b/src/core/NEON/kernels/arm_conv/depthwise/interleaves/a64_s8q_3x3_dot.cpp
index 5e4bf99120..3de4bdc1fb 100644
--- a/src/core/NEON/kernels/arm_conv/depthwise/interleaves/a64_s8q_3x3_dot.cpp
+++ b/src/core/NEON/kernels/arm_conv/depthwise/interleaves/a64_s8q_3x3_dot.cpp
@@ -42,7 +42,7 @@ size_t interleave_a64_s8q_3x3_dot::get_packed_size(const DepthwiseArgs &args)
{
// We store 7 vectors for every <vector_of_ints> of channels.
const unsigned int n = arm_gemm::roundup(
- arm_gemm::iceildiv((long unsigned int) args.input_channels,
+ arm_gemm::iceildiv((long unsigned int) args.input_channels * args.channel_multiplier,
get_vector_length<int32_t>(arm_gemm::VLType::None)), 4lu
);
return n * 7 * get_vector_length<int8_t>(arm_gemm::VLType::None);