Karol Herbst e2580bded2 Use CL_KERNEL_WORK_GROUP_SIZE more often (#2435) ...

Drivers _may_ choose to advertise values for
`CL_DEVICE_MAX_WORK_GROUP_SIZE` or `CL_DEVICE_MAX_WORK_ITEM_SIZES` that
kernels without a `reqd_work_group_size` are not able to be launched
with.

The CTS should therefore make sure that the local_size passed to
`clEnqueueNDRangeKernel` does not exceed `CL_KERNEL_WORK_GROUP_SIZE`

This fixes it up in two places I've noticed this not happening.

2025-08-05 14:40:11 -07:00

..

CMakeLists.txt

Queue profiling and Device timebase equality test (#1150)

2024-10-01 09:25:16 -07:00

copy.cpp

[NFC] Migrate profiling suite to new registration framework (#2187)

2024-12-17 09:57:27 -08:00

execute_multipass.cpp

Use CL_KERNEL_WORK_GROUP_SIZE more often (#2435)

2025-08-05 14:40:11 -07:00

execute.cpp

[NFC] Migrate profiling suite to new registration framework (#2187)

2024-12-17 09:57:27 -08:00

main.cpp

[NFC] Migrate profiling suite to new registration framework (#2187)

2024-12-17 09:57:27 -08:00

procs.h

[NFC] Migrate profiling suite to new registration framework (#2187)

2024-12-17 09:57:27 -08:00

profiling_timebase.cpp

[NFC] Migrate profiling suite to new registration framework (#2187)

2024-12-17 09:57:27 -08:00

readArray.cpp

Delete read_array_half and write_array_half tests (#2212)

2025-01-07 10:00:12 -08:00

readImage.cpp

[NFC] Migrate profiling suite to new registration framework (#2187)

2024-12-17 09:57:27 -08:00

writeArray.cpp

Delete read_array_half and write_array_half tests (#2212)

2025-01-07 10:00:12 -08:00

writeImage.cpp

[NFC] Migrate profiling suite to new registration framework (#2187)

2024-12-17 09:57:27 -08:00