intel/compute-runtime
1
0
Fork 0
mirror of https://github.com/intel/compute-runtime.git synced 2025-12-30 09:58:55 +08:00
Code Issues Packages Projects Releases Wiki Activity
18,423 Commits 42 Branches 655 Tags
b11322332ca95799bc48ce3560f67469504ffbd7
Commit Graph

2577 Commits

Author SHA1 Message Date
Szymon Morek
b11322332c performance: enable staging write for cl buffers 
Related-To: NEO-13529

Also, add size threshold on iGPU on Linux,
and disable staging if imported host ptr could
be reused

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
 2025-02-05 09:41:17 +01:00
Mateusz Jablonski
756c540275 refactor: remove pre-gen9 code 
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
 2025-02-04 12:29:05 +01:00
Mateusz Jablonski
0a068ce96a fix: add pipe control before scratch register write 
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
 2025-02-03 18:13:18 +01:00
Michał Pryba
75bc74089b refactor: adjust file names after pre-gen12 removal 2/3 
Related-To: NEO-12681
Signed-off-by: Michał Pryba <michal.pryba@intel.com>
 2025-02-03 15:31:51 +01:00
Compute-Runtime-Validation
d23249b061 Revert "fix: Set vmbind user fence when makeMemoryResident" 
This reverts commit 80dc4fb43a.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
 2025-01-31 11:36:29 +01:00
Kamil Kopryk
ef896cc799 refactor: introduce ImageSurfaceState helper class 
Moved global functions to the ImageSurfaceStateHelper class,
with declarations in the header file and definitions in the base .inl
file.
This change reduces compilation time by:
- removing unnecessary includes from the header file
- adding explicit template instantiations, which are faster than
implicit template instantiations.

Additionally, the image_skl_and_later.inl file has been removed as it
is no longer needed, and its implementation has been moved to the base .inl

Related-To: NEO-12149

Signed-off-by: Kamil Kopryk <kamil.kopryk@intel.com>
 2025-01-30 19:20:31 +01:00
Radoslaw Jablonski
30f552cd7d fix: resolve clang compilation issue on Windows 
A header was missing, making compilation using clang 18.1.8 fail.

Signed-off-by: Radoslaw Jablonski <radoslaw.jablonski@intel.com>
 2025-01-29 12:23:00 +01:00
Maciej Bielski
a8779c2387 fix: report ZE_MEMORY_ACCESS_CAP_FLAG_CONCURRENT correctly 
At the moment the capability is returned only based on the value
returned by the `productHelper`, which is too liberal. The capability
must also consider the support reported by `memoryManager`. Only then
the support reported is aligned with actual logic of handling
USM-allocations.

Related-To: NEO-10040
Signed-off-by: Maciej Bielski <maciej.bielski@intel.com>
 2025-01-29 00:17:38 +01:00
Chandio, Bibrak Qamar
80dc4fb43a fix: Set vmbind user fence when makeMemoryResident 
Related-To: NEO-11977, GSD-10293

Signed-off-by: Chandio, Bibrak Qamar <bibrak.qamar.chandio@intel.com>
 2025-01-28 22:04:37 +01:00
Chodor, Jaroslaw
574fe9fb29 refactor: Pass whole kernel descriptor to barrier programming 
Signed-off-by: Chodor, Jaroslaw <jaroslaw.chodor@intel.com>
 2025-01-27 01:26:49 +01:00
Dominik Dabek
bebeef0e88 feature: enable usm reuse cleaner 
Keep disabled in ULTs, except multi thread tests.

Related-To: NEO-13425

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
 2025-01-25 00:38:04 +01:00
Szymon Morek
d7c8b42d01 Revert "performance: enable staging write for cl buffers" 
This reverts commit a9a2f7577f.

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
 2025-01-24 23:41:12 +01:00
Szymon Morek
101c2fb071 fix: stop ULLS on main BCS when internal context used 
Related-To: NEO-13464

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
 2025-01-23 14:33:14 +01:00
Bartosz Dunajski
c75bcba1e6 fix: ensure profiling enabled for OCL Blit Split path 
Related-To: NEO-13842

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
 2025-01-23 13:59:15 +01:00
Michał Pryba
53831f61df refactor: adjust file names after pre-gen12 removal 1/3 
Related-To: NEO-12681
Signed-off-by: Michał Pryba <michal.pryba@intel.com>
 2025-01-22 22:47:24 +01:00
Bartosz Dunajski
c2dbdb6797 refactor: move blit post sync data to BlitProperties 
Related-To: NEO-13003

Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
 2025-01-22 18:35:21 +01:00
Damian Tomczak
4af92c20e4 feature: enable heapless for enqueueCopyBufferToImage 
Related-to: NEO-12737

Signed-off-by: Damian Tomczak <damian.tomczak@intel.com>
 2025-01-22 12:06:16 +01:00
Kamil Kopryk
6926582915 feature: use bindless samplers in heapless ocl 
Related-To: NEO-12741
Signed-off-by: Kamil Kopryk <kamil.kopryk@intel.com>
 2025-01-22 11:49:48 +01:00
Kamil Kopryk
47fc1ce34b refactor: use designated initializer for EncodeWalkerArgs 
also make EncodeWalkerArgs an aggregate

Signed-off-by: Kamil Kopryk <kamil.kopryk@intel.com>
 2025-01-22 11:44:57 +01:00
Zbigniew Zdanowicz
1e89058232 refactor: change data fields for encoder of additional walker fields 
Signed-off-by: Zbigniew Zdanowicz <zbigniew.zdanowicz@intel.com>
 2025-01-21 16:49:38 +01:00
Dominik Dabek
3f646839ca fix: usm reuse cleaning unused allocations 
mechanism for freeing allocations saved for reuse that have not been
used in a given time

Related-To: NEO-13425

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
 2025-01-21 14:23:19 +01:00
Zbigniew Zdanowicz
897c890d03 refactor: add thread group count parameter to implicit scaling functions 
Signed-off-by: Zbigniew Zdanowicz <zbigniew.zdanowicz@intel.com>
 2025-01-20 13:12:45 +01:00
Lukasz Jobczyk
af0bb7964a performance: Set dispatch all for small TG 
Resolves: NEO-11814

Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
 2025-01-20 08:49:59 +01:00
Zbigniew Zdanowicz
446f9c1964 refactor: change interfaces of walker region encoder 
Signed-off-by: Zbigniew Zdanowicz <zbigniew.zdanowicz@intel.com>
 2025-01-17 19:18:24 +01:00
Szymon Morek
ac78c07899 refactor: unify duplicated code for staging images 
Pass command type instead of maintaining
two methods for reads and writes

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
 2025-01-17 09:52:35 +01:00
Jaroslaw Warchulski
4dd1c80362 refactor: use enum class instead of enum for DeviceHierarchyMode 
Related-To: GSD-9560
Signed-off-by: Jaroslaw Warchulski <jaroslaw.warchulski@intel.com>
 2025-01-16 15:49:44 +01:00
Kamil Kopryk
d263fda1ad feature: patch 64bit addresses for bindless images in heapless OCL 
Related-To: NEO-12741
Signed-off-by: Kamil Kopryk <kamil.kopryk@intel.com>
 2025-01-16 15:24:02 +01:00
Szymon Morek
a9a2f7577f performance: enable staging write for cl buffers 
Related-To: NEO-13529

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
 2025-01-16 12:04:28 +01:00
Krzysztof Sprzaczkowski
a17745532c performance: Move preemption allocation init to the first submission 
Related-To: NEO-12323
Signed-off-by: Krzysztof Sprzaczkowski <krzysztof.sprzaczkowski@intel.com>
 2025-01-15 20:22:50 +01:00
Szymon Morek
0b4fe9a0df performance: add staging transfers for cl buffers 
Related-To: NEO-13529

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
 2025-01-15 12:24:43 +01:00
Compute-Runtime-Validation
af031ee0e3 Revert "performance: align structures for 64-bit platforms" 
This reverts commit 9f07f56f7f.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
 2025-01-15 09:02:01 +01:00
Jaroslaw Warchulski
6a7bee67e9 refactor: unify device hierarchy mode 
Related-To: GSD-9560
Signed-off-by: Jaroslaw Warchulski <jaroslaw.warchulski@intel.com>
 2025-01-14 21:08:44 +01:00
Damian Tomczak
b859a3b35b refactor: remove usesBindfulAddressingForBuffers function 
Related-to: NEO-12737

Signed-off-by: Damian Tomczak <damian.tomczak@intel.com>
 2025-01-14 10:45:04 +01:00
Jaroslaw Warchulski
a180afa2e6 fix: respect affinity mask with COMBINED hierarchy in OCL 
Related-To: GSD-9560
Signed-off-by: Jaroslaw Warchulski <jaroslaw.warchulski@intel.com>
 2025-01-13 15:38:39 +01:00
Jack Myers
7f9fadc314 fix: regression caused by tbx fault mngr 
Addresses regressions from the reverted merge
of the tbx fault manager for host memory.

Recursive locking of mutex caused deadlock.

To fix, separate tbx fault data from base
cpu fault data, allowing separate mutexes
for each, eliminating recursive locks on
the same mutex.

By separating, we also help ensure that tbx-related
changes don't affect the original cpu fault manager code
paths.

As an added safe guard preventing critical regressions
and avoiding another auto-revert, the tbx fault manager
is hidden behind a new debug flag which is disabled by default.

Related-To: NEO-12268
Signed-off-by: Jack Myers <jack.myers@intel.com>
 2025-01-09 07:48:53 +01:00
Semenov Herman (Семенов Герман)
9f07f56f7f performance: align structures for 64-bit platforms 
Signed-off-by: Semenov Herman (Семенов Герман) <GermanAizek@yandex.ru>
 2025-01-09 06:03:39 +01:00
Compute-Runtime-Validation
58707e0cf7 Revert "refactor: migration to std::make_unique C++17" 
This reverts commit 4cf685cb7d.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
 2025-01-09 01:41:36 +01:00
Semenov Herman (Семенов Герман)
4cf685cb7d refactor: migration to std::make_unique C++17 
Signed-off-by: Semenov Herman (Семенов Герман) <GermanAizek@yandex.ru>
 2025-01-08 17:20:00 +01:00
Szymon Morek
bd5ca4fed5 fix: flush ocl queue after staging transfer 
This is required for scenarios without direct submission
Otherwise, staging transfers are enqueued but not actually
flushed. If that happens, subsequent staging transfers can't
reuse previous staging buffers. It results in increased memory usage.

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
 2025-01-08 14:02:33 +01:00
Szymon Morek
2815d4167d fix: don't use staging for map allocations 
Related-To: NEO-13572

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
 2025-01-07 13:30:11 +01:00
Mateusz Jablonski
bb1a125f0c feature: add support for Panther Lake platform 
Related-To: NEO-12803

Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
 2025-01-07 11:39:56 +01:00
Mateusz Hoppe
fc4ff598c1 refactor: move heapless related inline data programming 
Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
 2024-12-30 18:40:48 +01:00
Szymon Morek
548f93c55a Revert "performance: enable staging read for images" 
This reverts commit 6aeb8dffa0.

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
 2024-12-23 17:42:12 +01:00
Szymon Morek
1135c1033e fix: set staging buffer writeable for aub and tbx 
Signed-off-by: Szymon Morek <szymon.morek@intel.com>
 2024-12-23 13:35:01 +01:00
Szymon Morek
6aeb8dffa0 performance: enable staging read for images 
Related-To: NEO-12968

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
 2024-12-20 15:02:28 +01:00
Compute-Runtime-Validation
124e755b9d Revert "fix: regression caused by tbx fault mngr" 
This reverts commit 9a14fe2478.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
 2024-12-19 17:35:03 +01:00
Jack Myers
9a14fe2478 fix: regression caused by tbx fault mngr 
Addresses regressions from the reverted merge
of the tbx fault manager for host memory.

This fixes attempts by the tbx fault manager
to protect/unprotect host buffer memory, even
if the host ptr was not driver-allocated.

In the case of the smoke test that triggered
the critical regression, clCreateBuffer was
called with the CL_MEM_USE_HOST_PTR flag.
The subsequent `mprotect` calls on the
provided host ptr then failed.

Related-To: NEO-12268
Signed-off-by: Jack Myers <jack.myers@intel.com>
 2024-12-18 23:16:36 +01:00
Szymon Morek
c5d541dd81 fix: initialize pointers to events 
Signed-off-by: Szymon Morek <szymon.morek@intel.com>
 2024-12-18 16:25:56 +01:00
Chodor, Jaroslaw
31f0eef4be fix: Use consistently proper size of implicit args buffer 
Fixes discrepancy of implicit args buffer allocation and programming
in OCL path

Resolves: NEO-12316

Signed-off-by: Chodor, Jaroslaw <jaroslaw.chodor@intel.com>
 2024-12-16 18:07:22 +01:00
Dominik Dabek
e61d04a881 fix(ocl): track buffer pool count per device 
Track amount of created buffer pools per device. Do not allocate extra
pools if limit is reached. New contexts will have pooling disabled if
limit is reached on device.

Related-To: NEO-13461

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
 2024-12-16 11:38:05 +01:00