diff options
| author | Giorgio Arena <giorgio.arena@arm.com> | 2018-08-23 12:00:02 +0100 |
|---|---|---|
| committer | Anthony Barbier <anthony.barbier@arm.com> | 2018-11-02 16:54:54 +0000 |
| commit | 226e4b92b191491ffa57ede66eba1d5d6fcf3b76 (patch) | |
| tree | 334705a1e743e3465400208d582cf0b25bf950fa /src/runtime/CL/tuners | |
| parent | 35aea3776449557c438e264bae7af5b1fe0e5ff6 (diff) | |
| download | ComputeLibrary-226e4b92b191491ffa57ede66eba1d5d6fcf3b76.tar.gz | |
COMPMID-1470 Add auto-init of the output in NECol2im
The output of NECol2Im is already auto-initialized.
This patch is about calling ShapeCalculator instead of computing the shape inside the kernel, adding validate_and_configure_window, and standardize the way convolved dims are passed (now NEON uses Size2D, while CL passes a pair of uint values: using Size2D for both implementations)
Change-Id: I795696e1b6532f57847c3186c1b532c09f5a25da
Reviewed-on: https://eu-gerrit-1.euhpc.arm.com/145345
Tested-by: Jenkins <bsgcomp@arm.com>
Reviewed-by: Michele DiGiorgio <michele.digiorgio@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Diffstat (limited to 'src/runtime/CL/tuners')
| -rw-r--r-- | src/runtime/CL/tuners/BifrostTuner.cpp | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/src/runtime/CL/tuners/BifrostTuner.cpp b/src/runtime/CL/tuners/BifrostTuner.cpp index 2d52f3392e..59d73b4e79 100644 --- a/src/runtime/CL/tuners/BifrostTuner.cpp +++ b/src/runtime/CL/tuners/BifrostTuner.cpp @@ -134,7 +134,7 @@ void tune_col2im_kernel(CLCol2ImKernel &k) // via exhaustive autotuning over 30 representative tensor shapes. if(gpu_target_is_in(gpu_target, GPUTarget::G71, GPUTarget::G72, GPUTarget::G51, GPUTarget::G51BIG, GPUTarget::G51LIT, GPUTarget::G76)) { - if((k._convolved_dims.first == 7) || (k._convolved_dims.first == 14)) + if((k._convolved_dims.width == 7) || (k._convolved_dims.width == 14)) { lws_hint = cl::NDRange(1, 7, 1); } |