diff options
author | Gunes Bayir <gunes.bayir@arm.com> | 2024-02-12 21:32:51 +0000 |
---|---|---|
committer | Gunes Bayir <gunes.bayir@arm.com> | 2024-02-21 10:36:22 +0000 |
commit | ef637398a8c2060e15de438020c53331da8bd6dd (patch) | |
tree | b1a1738736c9b6b49e76767e44bf4b77bf732876 /src/core/NEON/kernels/arm_gemm/gemv_pretransposed.hpp | |
parent | 0a48c4c83b598991b4d4235f870c24d9e6634b20 (diff) | |
download | ComputeLibrary-ef637398a8c2060e15de438020c53331da8bd6dd.tar.gz |
Integrate new pretranspose_b_array with extra fused transpose of B
This patch fuses the transposition taking place in Acl with the transformations done in arm_gemm (called pretranspose_b_array) if the underlying kernel and transform supports it. This should improve start-up time (as it's for constant Rhs matrices) and memory footprint. The transformations in arm_gemm are kernel specific. The Rhs matrix is transformed into certain layouts to improve the performance.
Resolves: COMPMID-6595
Change-Id: Id2932dd966e59f903c279417bebcea83d9a42464
Signed-off-by: Gunes Bayir <gunes.bayir@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/11144
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_gemm/gemv_pretransposed.hpp')
-rw-r--r-- | src/core/NEON/kernels/arm_gemm/gemv_pretransposed.hpp | 8 |
1 files changed, 5 insertions, 3 deletions
diff --git a/src/core/NEON/kernels/arm_gemm/gemv_pretransposed.hpp b/src/core/NEON/kernels/arm_gemm/gemv_pretransposed.hpp index 86b33d081f..f70fc98572 100644 --- a/src/core/NEON/kernels/arm_gemm/gemv_pretransposed.hpp +++ b/src/core/NEON/kernels/arm_gemm/gemv_pretransposed.hpp @@ -1,5 +1,5 @@ /* - * Copyright (c) 2017-2022 Arm Limited. + * Copyright (c) 2017-2022, 2024 Arm Limited. * * SPDX-License-Identifier: MIT * @@ -215,7 +215,9 @@ public: } } - void pretranspose_B_array(void *buffer, const To *B, const int ldb, const int B_multi_stride) override { + void pretranspose_B_array(void *buffer, const To *B, const int ldb, const int B_multi_stride, bool transposed) override { + assert(!transposed); + requantize_bias(buffer, B, ldb, B_multi_stride); // The actual transposed buffer goes after the column sums (if any) @@ -225,7 +227,7 @@ public: strategy strat(_args._ci); for (unsigned int multi=0; multi<_args._nmulti; multi++) { - strat.transforms.PrepareB(B_buffer + (multi * _buffer_per_multi), B + (multi * B_multi_stride), ldb, 0, _args._Nsize, 0, _args._Ksize); + strat.transforms.PrepareB(B_buffer + (multi * _buffer_per_multi), B + (multi * B_multi_stride), ldb, 0, _args._Nsize, 0, _args._Ksize, false); } _B_pretransposed = B_buffer; |