commit a668f9f8a4eab405df0fe8dd58e7d9425bcf9640
author Jonathan Deakin <jonathan.deakin@arm.com> Wed Jan 24 09:15:38 2024 +0000
committer Radu Salavat <radu.salavat@arm.com> Mon Apr 15 13:52:31 2024 +0000
tree db16e6af9289897557a58755b88d2c337dcb8650
parent 34bdffb288d6367cb6dca652ebed60c450854039

Add s8f32 kernels and dynamic QuantizationInfo

- Add support for QASYMM_SIGNED*QASYMM8_SIGNED->F32 in
  CpuGemmLowpMatrixMultiplyCore
- Add s8f32 kernel using existing s8->s32 kernels with a new
  DequantizeFloat OutputStage, the structure is similar to Requantize32
  but the opposite way around.
- Add SME s8f32 kernels with integrated support for DequantizeFloat.
- Add scale to CpuGemmLowpOffsetContributionKernel.
- Add virtual dequantize scale to gemm_common, only implemented for
  gemm_interleaved.
- Update year to 2024 in generate_build_files.
- Add dynamic flag to QuantizationInfo which signals to operators that
  it can change after configuration
- Add support for dynamic quantization in NEGEMMLowpMatrixMultiplyCore
- Add dynamic quantization fixture by extending
  GEMMLowpGenericMatrixMultiplyCoreValidationFixture
- Add GEMMLowpDequantizedMatrixMultiplyValidationFixture
- Store k (number of cols of A) rather than k_offset in the offset
  contribution kernels so that we can recompute it when the other
  offsets change

relates to: ONCPUML-1444 MLINFSW-439

Co-authored-by: Milos Puzovic <Milos.Puzovic@arm.com>
Co-authored-by: David Mansell <David.Mansell@arm.com>
Change-Id: I58a3acf2c09289a303e52eea6b336a696a5bc8da
Signed-off-by: Jonathan Deakin <jonathan.deakin@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/11022
Reviewed-by: Gunes Bayir <gunes.bayir@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@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 497da8e..2c3621c 100644
--- a/filelist.json
+++ b/filelist.json
@@ -1598,6 +1598,7 @@
               "src/core/NEON/kernels/arm_gemm/gemm_bf16bf16.cpp",
               "src/core/NEON/kernels/arm_gemm/gemm_int16.cpp",
               "src/core/NEON/kernels/arm_gemm/gemm_int8.cpp",
+              "src/core/NEON/kernels/arm_gemm/gemm_s8fp32.cpp",
               "src/core/NEON/kernels/arm_gemm/gemm_qint8.cpp",
               "src/core/NEON/kernels/arm_gemm/gemm_quint8.cpp",
               "src/core/NEON/kernels/arm_gemm/gemm_uint16.cpp",
@@ -1723,6 +1724,9 @@
               "src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_s8q_mopa_1VLx4VL/generic.cpp",
               "src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_s8q_mopa_2VLx2VL/generic.cpp",
               "src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_s8q_mopa_4VLx1VL/generic.cpp",
+              "src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_s8qfp32_mopa_1VLx4VL/generic.cpp",
+              "src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_s8qfp32_mopa_2VLx2VL/generic.cpp",
+              "src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_s8qfp32_mopa_4VLx1VL/generic.cpp",
               "src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_s8s32_mopa_1VLx4VL/generic.cpp",
               "src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_s8s32_mopa_2VLx2VL/generic.cpp",
               "src/core/NEON/kernels/arm_gemm/kernels/sme2_interleaved_nomerge_s8s32_mopa_4VLx1VL/generic.cpp",