aboutsummaryrefslogtreecommitdiff
path: root/src/core/NEON/kernels/arm_conv/depthwise/interleaves/sve_s8q_3x3_dot.cpp
diff options
context:
space:
mode:
authorMichael Tyler <michael.tyler@arm.com>2023-04-12 17:43:17 +0100
committermichael.tyler <michael.tyler@arm.com>2023-06-05 15:57:58 +0000
commit74921eee924625426429044decefe3673561b174 (patch)
tree654da1a95e3d42d6af8ad1ff27bb40d77b1fd8c5 /src/core/NEON/kernels/arm_conv/depthwise/interleaves/sve_s8q_3x3_dot.cpp
parentdf5d9878008be9b60586df97ebfff197abb5195e (diff)
downloadComputeLibrary-74921eee924625426429044decefe3673561b174.tar.gz
Update CPU kernel implementations and guard directives
Resolves COMPMID-6023 Change-Id: I868975d14c4f98af6716726feda22405a6a4c891 Signed-off-by: Michael Tyler <michael.tyler@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/9686 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Viet-Hoa Do <viet-hoa.do@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Benchmark: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/core/NEON/kernels/arm_conv/depthwise/interleaves/sve_s8q_3x3_dot.cpp')
-rw-r--r--src/core/NEON/kernels/arm_conv/depthwise/interleaves/sve_s8q_3x3_dot.cpp5
1 files changed, 2 insertions, 3 deletions
diff --git a/src/core/NEON/kernels/arm_conv/depthwise/interleaves/sve_s8q_3x3_dot.cpp b/src/core/NEON/kernels/arm_conv/depthwise/interleaves/sve_s8q_3x3_dot.cpp
index 0cf8044733..3a4999296a 100644
--- a/src/core/NEON/kernels/arm_conv/depthwise/interleaves/sve_s8q_3x3_dot.cpp
+++ b/src/core/NEON/kernels/arm_conv/depthwise/interleaves/sve_s8q_3x3_dot.cpp
@@ -25,8 +25,8 @@
#if defined(ARM_COMPUTE_ENABLE_SVE)
#include "arm_gemm.hpp"
-#include "src/core/NEON/kernels/arm_gemm/utils.hpp"
-#include "src/core/NEON/kernels/assembly/depthwise.hpp"
+#include "utils.hpp"
+#include "depthwise.hpp"
#include <cstdint>
namespace arm_conv {
@@ -76,7 +76,6 @@ void interleave_sve_s8q_3x3_dot::pack_parameters(unsigned int n_channels, void *
"cbz %x[bias], 1f\n"
"ptrue p8.s\n"
"1:" // No bias
-
"2:" // Loop
"cntp x20, p2, p1.s\n"
"whilelt p0.b, XZR, x20\n"