diff options
| author | Michalis Spyrou <michalis.spyrou@arm.com> | 2020-05-21 15:02:36 +0100 |
|---|---|---|
| committer | Georgios Pinitas <georgios.pinitas@arm.com> | 2020-06-16 11:42:09 +0000 |
| commit | bcd2352d7fd99a2f6aab220fa0c3b3f3119a1a4c (patch) | |
| tree | a3e1880071bca828b1c58be71805ccce4b205e53 /arm_compute/runtime/OMP | |
| parent | eae658453199d67a41deccbeb78e55b8eea9e966 (diff) | |
| download | ComputeLibrary-bcd2352d7fd99a2f6aab220fa0c3b3f3119a1a4c.tar.gz | |
COMPMID-3391: Implement Async interfaces
Change-Id: I8168cea5056ff48a0253ebb8c88ea549a3ea69a2
Signed-off-by: Michalis Spyrou <michalis.spyrou@arm.com>
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/3335
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Diffstat (limited to 'arm_compute/runtime/OMP')
| -rw-r--r-- | arm_compute/runtime/OMP/OMPScheduler.h | 15 |
1 files changed, 14 insertions, 1 deletions
diff --git a/arm_compute/runtime/OMP/OMPScheduler.h b/arm_compute/runtime/OMP/OMPScheduler.h index ed00833a9c..8ed1705a97 100644 --- a/arm_compute/runtime/OMP/OMPScheduler.h +++ b/arm_compute/runtime/OMP/OMPScheduler.h @@ -1,5 +1,5 @@ /* - * Copyright (c) 2017-2019 ARM Limited. + * Copyright (c) 2017-2020 ARM Limited. * * SPDX-License-Identifier: MIT * @@ -55,6 +55,19 @@ public: */ void schedule(ICPPKernel *kernel, const Hints &hints) override; + /** Multithread the execution of the passed kernel if possible. + * + * The kernel will run on a single thread if any of these conditions is true: + * - ICPPKernel::is_parallelisable() returns false + * - The scheduler has been initialized with only one thread. + * + * @param[in] kernel Kernel to execute. + * @param[in] hints Hints for the scheduler. + * @param[in] inputs Vector containing the input tensors. + * @param[in] outputs Vector containing the output tensors. + */ + void schedule_op(ICPPKernel *kernel, const Hints &hints, std::vector<InputOperatorTensors *> &inputs, std::vector<OutputOperatorTensors *> &outputs) override; + protected: /** Execute all the passed workloads * |