commit 6e09e1404c635d948cf20eb6b4b5747dfb6656f2
author Murray Kornelsen <murray.kornelsen@mail.mcgill.ca> Wed Jul 13 21:40:26 2022 -0400
committer Pablo Marquez Tello <pablo.tello@arm.com> Wed Sep 14 06:48:39 2022 +0000
tree 006199bd21b8a1330e1f1c86be60084bfb466706
parent a4814e8394ffdd7e268614d54cc22e30648f48ff

INT8 Quantized MeanStdDevNorm (LayerNorm)

Implements LayerNorm for qasymm8 tensors.
Uses uint8x16 loads and stores.
Summation is performed in integer arithmetic (vpaddl)
Normalization is performed in float32 before requantizing back to int8.

Signed-off-by: Murray Kornelsen <murray.kornelsen@mail.mcgill.ca>
Change-Id: I2407c8b34717fb47adab98791bd76fb8a3c62f4a
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7922
Comments-Addressed: Pablo Marquez Tello <pablo.tello@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Viet-Hoa Do <viet-hoa.do@arm.com>
Reviewed-by: Pablo Marquez Tello <pablo.tello@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>

filelist.json

diff --git a/filelist.json b/filelist.json
index c218ed9..eb39915 100644
--- a/filelist.json
+++ b/filelist.json
@@ -1738,7 +1738,8 @@
         "neon":{
           "common":["src/cpu/kernels/meanstddevnorm/generic/neon/impl.cpp"],
           "fp32":["src/cpu/kernels/meanstddevnorm/generic/neon/fp32.cpp"],
-          "fp16":["src/cpu/kernels/meanstddevnorm/generic/neon/fp16.cpp"]
+          "fp16":["src/cpu/kernels/meanstddevnorm/generic/neon/fp16.cpp"],
+          "qasymm8":["src/cpu/kernels/meanstddevnorm/generic/neon/qasymm8.cpp"]
         }
         }
       },