diff options
| author | SiCong Li <sicong.li@arm.com> | 2023-04-06 16:30:18 +0100 |
|---|---|---|
| committer | SiCong Li <sicong.li@arm.com> | 2023-04-26 09:10:38 +0000 |
| commit | dba672cec878966e465bb476e896c8f75bbd9145 (patch) | |
| tree | fcc8df3dc3f3799a616d2a10d52dd9bfdf6d2e33 /src/runtime | |
| parent | 7fefac722568d997b4d9e136925e93c7abeb564a (diff) | |
| download | ComputeLibrary-dba672cec878966e465bb476e896c8f75bbd9145.tar.gz | |
Integrate multi-threaded pretranspose_B_array
This is required for the case where rhs (B) is dynamic and needs to be
pretransposed in every run.
In a multi-threaded setting, this means the previously single-threaded
pretranspose_B_array would become the bottleneck
Resolves COMPMID-5896
Signed-off-by: SiCong Li <sicong.li@arm.com>
Change-Id: Id508c46992188a0f76a505152931d4955d04c16d
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/9455
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Viet-Hoa Do <viet-hoa.do@arm.com>
Reviewed-by: Jakub Sujak <jakub.sujak@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Benchmark: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'src/runtime')
| -rw-r--r-- | src/runtime/IScheduler.cpp | 6 |
1 files changed, 3 insertions, 3 deletions
diff --git a/src/runtime/IScheduler.cpp b/src/runtime/IScheduler.cpp index 39f41555fa..436fd9ca16 100644 --- a/src/runtime/IScheduler.cpp +++ b/src/runtime/IScheduler.cpp @@ -1,5 +1,5 @@ /* - * Copyright (c) 2016-2022 Arm Limited. + * Copyright (c) 2016-2023 Arm Limited. * * SPDX-License-Identifier: MIT * @@ -139,7 +139,7 @@ void IScheduler::schedule_common(ICPPKernel *kernel, const Hints &hints, const W default: ARM_COMPUTE_ERROR("Unknown strategy"); } - // Make sure the smallest window is larger than minimim workload size + // Make sure the smallest window is larger than minimum workload size num_windows = adjust_num_of_windows(max_window, hints.split_dimension(), num_windows, *kernel, cpu_info()); std::vector<IScheduler::Workload> workloads(num_windows); @@ -178,7 +178,7 @@ void IScheduler::run_tagged_workloads(std::vector<Workload> &workloads, const ch std::size_t IScheduler::adjust_num_of_windows(const Window &window, std::size_t split_dimension, std::size_t init_num_windows, const ICPPKernel &kernel, const CPUInfo &cpu_info) { // Mitigation of the narrow split issue, which occurs when the split dimension is too small to split (hence "narrow"). - if(window.num_iterations(split_dimension) < init_num_windows ) + if(window.num_iterations(split_dimension) < init_num_windows) { auto recommended_split_dim = Window::DimX; for(std::size_t dims = Window::DimY; dims <= Window::DimW; ++dims) |