Mateusz Tabaka ff6ea62ce0 Fix local work size for conv kernel yxfb_yxio_b16 with fp16 (#11679) ...

convolution_gpu_yxfb_yxio_b16 for fp16 has hardcoded reqd_work_group_size
to (16, 1, 1). On devices where CL_DEVICE_MAX_WORK_GROUP_SIZE is 512
GetOptimalLocalWorkGroupSizes picks (16, 2, 1) for LWS.
That causes issues during clEnqueueNDRangeKernel since LWS doesn't match
with reqd_work_group_size in the kernel.

2022-05-23 15:59:37 +02:00

..

bindings

[ROIAlign-9] Extend nGraph Python API for operation "ROIAlign-9" (#11572)

2022-05-19 13:55:46 +02:00

cmake

Use system dependencies (#11419)

2022-04-05 04:47:22 +03:00

common

Meiyang/paddle generate proposals 2 (#11285)

2022-05-16 09:13:52 +08:00

core

[Eye-9] Reference and CPU implementation for Eye-9 (#11538)

2022-05-19 16:37:00 +03:00

frontends

ONNX node names as friendly names (#10532)

2022-05-17 13:47:52 +02:00

inference

Move RemoveConcatZeroDimInput and RemoveMultiSubGraphOpDanglingParams to CNNNetworkNGraphImpl ctor (#11547)

2022-05-19 10:50:22 +02:00

plugins

Fix local work size for conv kernel yxfb_yxio_b16 with fp16 (#11679)

2022-05-23 15:59:37 +02:00

tests

[GNA] Change Constants' precision in MVN decomposition (#11665)

2022-05-20 13:21:32 +02:00

tests_deprecated

[GNA] Moved PWL functional tests (#10731)

2022-04-05 11:43:24 +03:00

.clang-format

Move core components (#8869)

2021-11-27 18:53:12 +03:00

CMakeLists.txt

Use system dependencies (#11419)

2022-04-05 04:47:22 +03:00