Dominik Dabek 39cf653959 performance(ocl): cmd buffer prealloc per cmdqueue ...

Add mechanism to preallocate cmd buffer allocations in command stream
receiver reusable allocations list per command queue initialized.

This should limit additional allocations during hot loop.

Needs to be enabled in subsequent commits by setting product helper
method.

Related-To: NEO-8152

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>

2023-10-27 16:56:29 +02:00

..

definitions

performance: add new copy operations to state base address properties

2023-06-07 11:34:28 +02:00

aub_command_stream_receiver_hw_base.inl

feature: allow freeing memory in aubstream

2023-10-26 17:16:23 +02:00

aub_command_stream_receiver_hw_bdw_and_later.inl

style: correct variable naming

2022-05-16 15:02:15 +02:00

aub_command_stream_receiver_hw_xehp_and_later.inl

Move generic command stream receiver files to shared

2021-09-22 23:55:31 +02:00

aub_command_stream_receiver_hw.h

fix: capability to write memory chunk in aub/tbx mode

2023-10-19 19:13:11 +02:00

aub_command_stream_receiver.cpp

Revert "fix: if device hierarchy is flat then getSubDevicesCount return 1u"

2023-10-26 15:40:29 +02:00

aub_command_stream_receiver.h

Revert "fix: if device hierarchy is flat then getSubDevicesCount return 1u"

2023-10-26 15:40:29 +02:00

aub_subcapture_status.h

Reorganization directory structure [3/n]

2020-02-23 23:48:28 +01:00

CMakeLists.txt

Split the L0 BCS split into D2H and H2D

2023-02-13 14:17:39 +01:00

command_stream_receiver_hw_base.inl

fix: add unrecoverable in case of nullptr

2023-10-04 14:38:05 +02:00

command_stream_receiver_hw_bdw_and_later.inl

performance(ocl): flag to not dcFlush on no event

2023-07-18 15:38:54 +02:00

command_stream_receiver_hw_dg2_and_later.inl

fix: improve release helper

2023-05-15 14:30:15 +02:00

command_stream_receiver_hw_xehp_and_later.inl

fix: add l1 cache flush to barrier operations when required

2023-10-23 18:18:22 +02:00

command_stream_receiver_hw.h

refactor: remove not needed code

2023-09-27 14:35:49 +02:00

command_stream_receiver_simulated_common_hw_base.inl

fix: decanonize GPU VA during TBX memory read.

2023-07-26 19:44:19 +02:00

command_stream_receiver_simulated_common_hw_bdw_and_later.inl

Move generic command stream receiver files to shared

2021-09-22 23:55:31 +02:00

command_stream_receiver_simulated_common_hw_xehp_and_later.inl

Revert "fix: if device hierarchy is flat then getSubDevicesCount return 1u"

2023-10-26 15:40:29 +02:00

command_stream_receiver_simulated_common_hw.h

fix: fix compilation error in clang on Windows (2/n)

2023-10-24 15:59:06 +02:00

command_stream_receiver_simulated_hw.h

Revert "fix: if device hierarchy is flat then getSubDevicesCount return 1u"

2023-10-26 15:40:29 +02:00

command_stream_receiver_with_aub_dump.h

fix: capability to write memory chunk in aub/tbx mode

2023-10-19 19:13:11 +02:00

command_stream_receiver_with_aub_dump.inl

fix: capability to write memory chunk in aub/tbx mode

2023-10-19 19:13:11 +02:00

command_stream_receiver.cpp

performance(ocl): cmd buffer prealloc per cmdqueue

2023-10-27 16:56:29 +02:00

command_stream_receiver.h

performance(ocl): cmd buffer prealloc per cmdqueue

2023-10-27 16:56:29 +02:00

create_command_stream_impl.cpp

Cleanup includes 15

2022-12-23 12:36:51 +01:00

create_command_stream_impl.h

zesSysmanDeviceReset: Reinitialize device after device reset

2021-11-07 23:43:48 +01:00

csr_definitions.h

refactor: pass big parameters as reference instead of by value

2023-10-04 14:53:13 +02:00

csr_deps.cpp

fix make resident dependent multi root device tags

2023-02-14 09:58:00 +01:00

csr_deps.h

Cleanup includes 45

2023-01-26 11:12:15 +01:00

csr_properties_flags.h

Reduce binaries sizes 2/n

2022-12-08 16:14:38 +01:00

device_command_stream.h

Refactor of global factories

2021-05-28 19:09:24 +02:00

experimental_command_buffer.cpp

Cleanup includes 15

2022-12-23 12:36:51 +01:00

experimental_command_buffer.h

refactor: add NonCopyableOrMovableClass to classes that don't need copy ctors

2023-10-05 13:11:01 +02:00

experimental_command_buffer.inl

feature: add indirect semaphore mode

2023-09-12 13:15:51 +02:00

linear_stream.cpp

Cleanup headers

2022-05-18 11:42:06 +02:00

linear_stream.h

fix: commandContainer is nullptr in LinearStream for immediate cmdList

2023-07-24 15:06:18 +02:00

memory_compression_state.h

Add helper functions for memory compression to CSR

2020-12-21 17:43:03 +01:00

preemption_mode.h

Reorganization directory structure [3/n]

2020-02-23 23:48:28 +01:00

preemption_xehp_and_later.inl

refactor: remove not needed code

2023-09-27 14:35:49 +02:00

preemption.cpp

Cleanup includes 48

2023-02-10 17:23:02 +01:00

preemption.h

refactor: remove not needed code

2023-09-27 14:35:49 +02:00

preemption.inl

refactor: remove not needed code

2023-09-27 14:35:49 +02:00

queue_throttle.h

Reorganization directory structure [3/n]

2020-02-23 23:48:28 +01:00

scratch_space_controller_base.cpp

refactor: rename hw_helper files to gfx_core_helper files

2023-02-01 19:37:51 +01:00

scratch_space_controller_base.h

Use dedicated using type for TaskCount

2022-11-28 16:44:44 +01:00

scratch_space_controller_xehp_and_later.cpp

feature: bindless global allocator with scratch

2023-08-03 13:24:59 +02:00

scratch_space_controller_xehp_and_later.h

Use dedicated using type for TaskCount

2022-11-28 16:44:44 +01:00

scratch_space_controller.cpp

refactor: rename hw_helper files to gfx_core_helper files

2023-02-01 19:37:51 +01:00

scratch_space_controller.h

refactor: add NonCopyableOrMovableClass to classes that don't need copy ctors

2023-10-05 13:11:01 +02:00

stream_properties_extra.cpp

[perf] reduce host overhead in command list reset call

2023-04-05 11:29:39 +02:00

stream_properties.cpp

performance: add new copy operations to state base address properties

2023-06-07 11:34:28 +02:00

stream_properties.h

[perf] reduce host overhead in command list reset call

2023-04-05 11:29:39 +02:00

stream_property.h

[perf] simplify state transition for size properties

2023-04-04 10:41:36 +02:00

submission_status.h

Propagate exec buffer error to L0 API level on Xe HPC

2022-10-31 10:09:13 +01:00

submissions_aggregator.cpp

fix: fix compilation error in clang on Windows (2/n)

2023-10-24 15:59:06 +02:00

submissions_aggregator.h

fix: fix compilation error in clang on Windows (2/n)

2023-10-24 15:59:06 +02:00

tag_allocation_layout.h

Introduce barrier tracking mechanism

2023-02-06 14:29:23 +01:00

task_count_helper.h

Change task count type to uint64_t

2022-11-29 12:05:22 +01:00

tbx_command_stream_receiver_hw.h

fix: fix compilation error in clang on Windows (2/n)

2023-10-24 15:59:06 +02:00

tbx_command_stream_receiver_hw.inl

Revert "fix: if device hierarchy is flat then getSubDevicesCount return 1u"

2023-10-26 15:40:29 +02:00

tbx_command_stream_receiver_xehp_and_later.inl

Move generic command stream receiver files to shared

2021-09-22 23:55:31 +02:00

tbx_command_stream_receiver.cpp

Move tbx stream, tbx csr to shared

2021-05-31 14:35:32 +02:00

tbx_command_stream_receiver.h

Cleanup includes 42

2023-01-25 09:16:39 +01:00

tbx_stream.cpp

Move tbx stream, tbx csr to shared

2021-05-31 14:35:32 +02:00

thread_arbitration_policy.h

Cleanup headers

2022-05-18 11:42:06 +02:00

transfer_direction.h

Split the L0 BCS split into D2H and H2D

2023-02-13 14:17:39 +01:00

wait_status.h

Poll task count indefinitely on high throttle command queue

2022-03-25 10:06:16 +01:00