Mateusz Hoppe b03f625f03 feature: limit max LWS based on preferred number of workgroups per ss ...

- limit max LWS size when SLM and barriers are not used

Related-To: GSD-11112

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>

2025-05-16 20:28:00 +02:00

..

cl_local_work_size.cpp

feature: limit max LWS based on preferred number of workgroups per ss

2025-05-16 20:28:00 +02:00

cl_local_work_size.h

Cleanup includes 34

2023-01-17 14:42:04 +01:00

CMakeLists.txt

refactor: adjust file names after pre-gen12 removal 2/3

2025-02-03 15:31:51 +01:00

command_queue_hw_base.inl

feature: Add 3-level wait scheme with tpause intrinsic

2025-03-21 12:12:57 +01:00

command_queue_hw_xehp_and_later.inl

fix:pass bool to get CS size instead of reevaluate

2023-03-24 15:24:10 +01:00

command_queue_hw.h

performance: introduce staging read for cl_buffer

2025-03-11 08:19:28 +01:00

command_queue_staging.cpp

fix: properly pass info about 3D image

2025-04-03 16:33:57 +02:00

command_queue.cpp

fix: do not allow blit for depth image on arl

2025-05-13 15:56:47 +02:00

command_queue.h

refactor: Remove unused cmdq round robin engine assign

2025-05-09 11:03:57 +02:00

copy_engine_state.h

Revert "performance: align structures for 64-bit platforms"

2025-01-15 09:02:01 +01:00

cpu_data_transfer_handler.cpp

refactor: correct naming of enum class constants 8/n

2023-12-19 08:18:18 +01:00

csr_selection_args.cpp

Cleanup includes 43

2023-01-25 11:33:39 +01:00

csr_selection_args.h

Revert "performance: align structures for 64-bit platforms"

2025-01-15 09:02:01 +01:00

enqueue_barrier.h

Cleanup includes

2022-07-25 09:58:38 +02:00

enqueue_common.h

fix: program texture cache flush before BCS copy

2025-03-12 14:41:15 +01:00

enqueue_copy_buffer_rect.h

feature: force stateless for copy buffer rect

2025-03-24 14:14:44 +01:00

enqueue_copy_buffer_to_image.h

feature: enable heapless for enqueueCopyBufferToImage

2025-01-22 12:06:16 +01:00

enqueue_copy_buffer.h

feature: force stateless for copy buffer

2025-03-24 14:29:25 +01:00

enqueue_copy_image_to_buffer.h

feature: use heapless builtins for images

2024-10-25 16:42:27 +02:00

enqueue_copy_image.h

feature: use heapless builtins for images

2024-10-25 16:42:27 +02:00

enqueue_fill_buffer.h

feature: use heapless builtins in OCL if supported

2024-02-08 15:49:58 +01:00

enqueue_fill_image.h

feature: add FillImage1dBuffer built-in kernel

2025-04-10 14:04:13 +02:00

enqueue_kernel.h

refactor: remove not needed code

2023-09-12 08:51:58 +02:00

enqueue_marker.h

Cleanup includes

2022-07-25 09:58:38 +02:00

enqueue_migrate_mem_objects.h

Detect GPU hangs in blocking enqueue handler calls

2022-03-22 21:06:19 +01:00

enqueue_read_buffer_rect.h

feature: use heapless builtins in OCL if supported

2024-02-08 15:49:58 +01:00

enqueue_read_buffer.h

performance: introduce staging read for cl_buffer

2025-03-11 08:19:28 +01:00

enqueue_read_image.h

feature: additional fixes to enable feature and adjusting builtins types

2025-03-20 14:23:08 +01:00

enqueue_resource_barrier.h

Detect GPU hangs in blocking enqueue handler calls

2022-03-22 21:06:19 +01:00

enqueue_svm.h

fix: adjust svm builtins types

2025-03-24 13:55:48 +01:00

enqueue_write_buffer_rect.h

feature: use heapless builtins in OCL if supported

2024-02-08 15:49:58 +01:00

enqueue_write_buffer.h

performance: add staging transfers for cl buffers

2025-01-15 12:24:43 +01:00

enqueue_write_image.h

feature: additional fixes to enable feature and adjusting builtins types

2025-03-20 14:23:08 +01:00

finish.h

refactor: correct naming of enum class constants

2023-12-01 11:53:51 +01:00

flush.h

Copyright header update

2021-05-17 20:38:19 +02:00

gpgpu_walker_base.inl

Revert "performance: use RESOURCE_BARRIER as stalling barrier"

2025-05-16 20:19:06 +02:00

gpgpu_walker_xehp_and_later.inl

refactor: remove unused parameter

2025-05-15 12:23:27 +02:00

gpgpu_walker.h

refactor: add setupTimestampPacketFlushL3 function

2025-03-24 13:10:03 +01:00

hardware_interface_base.inl

refactor: remove unused kernel tunning

2025-05-08 14:18:21 +02:00

hardware_interface_xehp_and_later.inl

fix: Add debug toggle to disable flush L3 for host usm

2025-05-07 16:57:31 +02:00

hardware_interface.h

refactor: cleanup headers

2025-04-14 14:59:40 +02:00

resource_barrier.cpp

refactor: use reference in for-range loops

2023-10-03 18:56:53 +02:00

resource_barrier.h

fix: Add asserts to ensure NonCopyable and NonMovable 1/n

2025-02-18 09:41:20 +01:00