diff options
author | Joseph Dobson <joseph.dobson@arm.com> | 2020-02-11 19:32:11 +0000 |
---|---|---|
committer | Gian Marco Iodice <gianmarco.iodice@arm.com> | 2020-05-10 13:34:11 +0000 |
commit | 6f8b17dedb7b53b550e6210fd1c78c3a3e086271 (patch) | |
tree | 6b040314802ea84f1ae84dda235c1af808863346 /src/core/NEON/kernels/arm_gemm/gemm_native.hpp | |
parent | 2886c757389c0ccca20a8689daf8180a730ecbc9 (diff) | |
download | ComputeLibrary-6f8b17dedb7b53b550e6210fd1c78c3a3e086271.tar.gz |
[ONCPUML-7] arm_compute support for ND parallelism
Currently 1D ranges of work are specified by the scheduler
via two integers, start and end. This limit opportunities
for advance parallelism and scheduling
This patch expands the interfaces to allow for ND parallism.
`GemmCommon::get_window_size` now returns an `NDRange` specifying the work
in N-dimensions rather than with the single integer it used prior (1D)
Execute now takes an `NDCoordinate` which specifies an `NDRange` with a start
position for that work along with an `NDCoordinate` to specify the thread location
In addition to expanding the interface to enable this functionality,
we have added the capability to SGEMM when the number of threads is high
this has the effective of allowing a much greater degree of parallelism
where te problem dimension would previously have limited the number of threads.
Change-Id: I3e1a8b7276216627bec4ff6f24ac2147552ea9fb
Signed-off-by: Joseph Dobson <joseph.dobson@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/2962
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/core/NEON/kernels/arm_gemm/gemm_native.hpp')
-rw-r--r-- | src/core/NEON/kernels/arm_gemm/gemm_native.hpp | 18 |
1 files changed, 14 insertions, 4 deletions
diff --git a/src/core/NEON/kernels/arm_gemm/gemm_native.hpp b/src/core/NEON/kernels/arm_gemm/gemm_native.hpp index fe6ebef045..c2f742b5cf 100644 --- a/src/core/NEON/kernels/arm_gemm/gemm_native.hpp +++ b/src/core/NEON/kernels/arm_gemm/gemm_native.hpp @@ -1,5 +1,5 @@ /* - * Copyright (c) 2017-2019 Arm Limited. + * Copyright (c) 2017-2020 Arm Limited. * * SPDX-License-Identifier: MIT * @@ -87,8 +87,8 @@ public: _window_range(iceildiv(_Msize, strategy::out_height()), _nbatches, iceildiv(_Nsize, _n_block), _nmultis) { } // Window is amount per multi multiplied by total number of multis. - unsigned int get_window_size() const override { - return _window_range.total_size(); + ndrange_t get_window_size() const override { + return { _window_range.total_size(), 1u, 1u, 1u, 1u, 1u }; } // Native GEMMs can always be dynamically scheduled (whether requested or not) @@ -97,7 +97,7 @@ public: } // Actually execute the GEMM. - void execute(unsigned int start, unsigned int end, int) override { + void execute_1d(unsigned int start, unsigned int end, int) { #ifdef CYCLE_PROFILING profiler prof; #endif @@ -139,6 +139,16 @@ public: } } while (p.next_dim1()); } + + //Execute + void execute(const ndcoord_t& work_range, const ndcoord_t& thread_locator, int threadid) override { + UNUSED(thread_locator); + + const auto start = work_range.get_position(0); + const auto stop = work_range.get_position_end(0); + + execute_1d(start, stop, threadid); + } }; } // namespace arm_gemm |