diff options
author | Georgios Pinitas <georgios.pinitas@arm.com> | 2021-07-16 16:16:43 +0100 |
---|---|---|
committer | Georgios Pinitas <georgios.pinitas@arm.com> | 2021-07-22 02:25:50 +0000 |
commit | 4ee8b1599dbaf7634d25607fa5ac96ba3dc6b0f2 (patch) | |
tree | 2f8362d33cdad4212f4b96995681c68184c759e1 /src/core/NEON/kernels/arm_gemm/quantize_wrapper.hpp | |
parent | 59fd7a722e5bc7e85309d6200bc37a772721a719 (diff) | |
download | ComputeLibrary-4ee8b1599dbaf7634d25607fa5ac96ba3dc6b0f2.tar.gz |
Update GEMM assembly kernels
- Introduce Fp32 kernels with internal calculations in Bfloat16 when
fast_mode is enabled
- Improve kernel selection heuristics
Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com>
Change-Id: I68a9e7e862b6fd2721b46e0d7cc791091c4ab279
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5965
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/core/NEON/kernels/arm_gemm/quantize_wrapper.hpp')
-rw-r--r-- | src/core/NEON/kernels/arm_gemm/quantize_wrapper.hpp | 15 |
1 files changed, 14 insertions, 1 deletions
diff --git a/src/core/NEON/kernels/arm_gemm/quantize_wrapper.hpp b/src/core/NEON/kernels/arm_gemm/quantize_wrapper.hpp index fdb4f584d8..1e2a9acc1d 100644 --- a/src/core/NEON/kernels/arm_gemm/quantize_wrapper.hpp +++ b/src/core/NEON/kernels/arm_gemm/quantize_wrapper.hpp @@ -1,5 +1,5 @@ /* - * Copyright (c) 2019-2020 Arm Limited. + * Copyright (c) 2019-2021 Arm Limited. * * SPDX-License-Identifier: MIT * @@ -198,6 +198,19 @@ public: _params.bias = bias; _params.bias_multi_stride = bias_multi_stride; } + + GemmConfig get_config() override { + GemmConfig c = _subgemm->get_config(); + + std::string n = "quantize_wrapper["; + n.append(c.filter); + n.append("]"); + + c.method = GemmMethod::QUANTIZE_WRAPPER; + c.filter = n; + + return c; + } }; } // namespace arm_gemm |