diff options
| author | Gian Marco Iodice <gianmarco.iodice@arm.com> | 2019-03-15 10:13:05 +0000 |
|---|---|---|
| committer | Gian Marco Iodice <gianmarco.iodice@arm.com> | 2019-03-20 11:21:46 +0000 |
| commit | b0c5037d94ba7073ccabb0ebaff54db320f184c4 (patch) | |
| tree | 126f2332df60b6eff1e630b2585b2bd407501a20 /arm_compute/core/CL/ICLKernel.h | |
| parent | 5ed7b5bc98feb848874730c9bb9c30759e58d453 (diff) | |
| download | ComputeLibrary-b0c5037d94ba7073ccabb0ebaff54db320f184c4.tar.gz | |
COMPMID-2043: Add support for "dummy threads" in CLGEMMReshaped
Change-Id: I89403b97503fbb99f6a32f5d62b8c535ab26a7be
Signed-off-by: Gian Marco Iodice <gianmarco.iodice@arm.com>
Reviewed-on: https://review.mlplatform.org/c/877
Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'arm_compute/core/CL/ICLKernel.h')
| -rw-r--r-- | arm_compute/core/CL/ICLKernel.h | 14 |
1 files changed, 8 insertions, 6 deletions
diff --git a/arm_compute/core/CL/ICLKernel.h b/arm_compute/core/CL/ICLKernel.h index f5423584e1..590f8929cb 100644 --- a/arm_compute/core/CL/ICLKernel.h +++ b/arm_compute/core/CL/ICLKernel.h @@ -1,5 +1,5 @@ /* - * Copyright (c) 2016-2018 ARM Limited. + * Copyright (c) 2016-2019 ARM Limited. * * SPDX-License-Identifier: MIT * @@ -308,14 +308,16 @@ private: * * @note If kernel->kernel() is empty then the function will return without adding anything to the queue. * - * @param[in,out] queue OpenCL command queue. - * @param[in] kernel Kernel to enqueue - * @param[in] window Window the kernel has to process. - * @param[in] lws_hint Local workgroup size requested. Default is based on the device target. + * @param[in,out] queue OpenCL command queue. + * @param[in] kernel Kernel to enqueue + * @param[in] window Window the kernel has to process. + * @param[in] lws_hint (Optional) Local workgroup size requested. Default is based on the device target. + * @param[in] use_dummy_work_items (Optional) Use dummy work items in order to have two dimensional power of two NDRange. Default is false + * Note: it is kernel responsibility to check if the work-item is out-of-range * * @note If any dimension of the lws is greater than the global workgroup size then no lws will be passed. */ -void enqueue(cl::CommandQueue &queue, ICLKernel &kernel, const Window &window, const cl::NDRange &lws_hint = CLKernelLibrary::get().default_ndrange()); +void enqueue(cl::CommandQueue &queue, ICLKernel &kernel, const Window &window, const cl::NDRange &lws_hint = CLKernelLibrary::get().default_ndrange(), bool use_dummy_work_items = false); /** Add the passed array's parameters to the object's kernel's arguments starting from the index idx. * |