commit 94672fb2af6535adc6ea7fe8b8498580ad8cf3f4
author Georgios Pinitas <georgios.pinitas@arm.com> Wed Jan 22 18:36:27 2020 +0000
committer Georgios Pinitas <georgios.pinitas@arm.com> Fri Jan 31 11:36:14 2020 +0000
tree 189c68e7ff2c29d7800e09239da6d9a6364d5b05
parent 6a342648ae50beb8457871862f14fc9baef6b74f

COMPMID-3003: Integrate assembly kernels utilizing MMLA instruction.

MMLA is a matrix-multiply instruction introduced on armv8.6-A

Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com>
Change-Id: I572a54981d48f5a1e0e9e51102cb7ae28ad87806
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/2663
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Michalis Spyrou <michalis.spyrou@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>

Android.bp

diff --git a/Android.bp b/Android.bp
index 8ebe4c5..e7ad651 100644
--- a/Android.bp
+++ b/Android.bp
@@ -764,6 +764,8 @@
                 "src/core/NEON/kernels/arm_gemm/kernels/a64_hybrid_s8s32_dot_16x4/generic.cpp",
                 "src/core/NEON/kernels/arm_gemm/kernels/a64_hybrid_u8u32_dot_16x4/a55.cpp",
                 "src/core/NEON/kernels/arm_gemm/kernels/a64_hybrid_u8u32_dot_16x4/generic.cpp",
+                "src/core/NEON/kernels/arm_gemm/kernels/a64_interleaved_s8s32_mmla_12x8/generic.cpp",
+                "src/core/NEON/kernels/arm_gemm/kernels/a64_interleaved_u8u32_mmla_12x8/generic.cpp",
                 "src/core/NEON/kernels/arm_gemm/kernels/a64_native_fp32_mla_16x4/generic.cpp",
                 "src/core/NEON/kernels/arm_gemm/kernels/a64_sgemm_12x8/a53.cpp",
                 "src/core/NEON/kernels/arm_gemm/kernels/a64_sgemm_12x8/a55.cpp",
@@ -788,7 +790,9 @@
                 "src/core/NEON/kernels/arm_gemm/kernels/sve_interleaved_fp16_mla_3VLx8/generic.cpp",
                 "src/core/NEON/kernels/arm_gemm/kernels/sve_interleaved_fp32_mla_3VLx8/generic.cpp",
                 "src/core/NEON/kernels/arm_gemm/kernels/sve_interleaved_s8s32_dot_3VLx8/generic.cpp",
+                "src/core/NEON/kernels/arm_gemm/kernels/sve_interleaved_s8s32_mmla_3VLx8/generic.cpp",
                 "src/core/NEON/kernels/arm_gemm/kernels/sve_interleaved_u8u32_dot_3VLx8/generic.cpp",
+                "src/core/NEON/kernels/arm_gemm/kernels/sve_interleaved_u8u32_mmla_3VLx8/generic.cpp",
                 "src/core/NEON/kernels/arm_gemm/kernels/sve_native_fp16_mla_4VLx4/generic.cpp",
                 "src/core/NEON/kernels/arm_gemm/kernels/sve_native_fp32_mla_4VLx4/generic.cpp",
                 "src/core/NEON/kernels/arm_gemm/kernels/sve_native_s8s32_dot_4VLx4/generic.cpp",