intel/compute-runtime
1
0
Fork 0
mirror of https://github.com/intel/compute-runtime.git synced 2026-01-01 12:33:12 +08:00
Code Issues Packages Projects Releases Wiki Activity
21,422 Commits 43 Branches 655 Tags
2fbb12b545ff324a6396be2b65839ef31774ece7
Commit Graph

9017 Commits

Author SHA1 Message Date
Filip Hazubski
2fbb12b545 feature: Add debug toggle OverrideSamplerArbitrationControl 
Resolves: NEO-16608

Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
 2025-11-27 09:53:06 +01:00
Mateusz Jablonski
9db471bf13 test: add test for xe drm support 
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
 2025-11-27 09:51:42 +01:00
Mateusz Jablonski
1bfca607c9 test: add test for config in aub file name 
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
 2025-11-27 07:15:22 +01:00
Compute-Runtime-Validation
96457f879b Revert "fix: BMG returns non-zero maxClockRate in zeDeviceGetMemoryProperties" 
This reverts commit 175c8a821f.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
 2025-11-27 03:11:39 +01:00
Compute-Runtime-Validation
e9f1e05f31 Revert "fix: usm pool alignment check, use host ptr" 
This reverts commit b0530e13e2.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
 2025-11-27 02:12:31 +01:00
Szymon Morek
9892a42e6e performance: copy D2H with staging buffers 
Related-To: NEO-16791

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
 2025-11-26 18:02:11 +01:00
Szymon Morek
1ebfab3beb performance: enable timestamp caching on L0 
Related-To: NEO-14360

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
 2025-11-26 17:10:54 +01:00
Dominik Dabek
b0530e13e2 fix: usm pool alignment check, use host ptr 
Any alignments below pool alignment can be accepted.

Do not pool if host ptr is passed from application.

Adjust ULTs that previously did not use pooling due to alignment.

Related-To: NEO-16084

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
 2025-11-26 16:52:33 +01:00
Mateusz Jablonski
73a6b75f57 refactor: placeholder for device handle in gmm client context 
Related-To: NEO-11080

Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
 2025-11-26 15:28:29 +01:00
Michal Mrozek
af43bd6236 performance: add debug flag to override node ordinal for ccs 
Signed-off-by: Michal Mrozek <michal.mrozek@intel.com>
 2025-11-26 13:04:57 +01:00
Jack Myers
175c8a821f fix: BMG returns non-zero maxClockRate in zeDeviceGetMemoryProperties 
Implemented getDeviceMemoryMaxClkRate() override for BMG Linux to query
actual memory clock rate from DRM driver instead of returning 0.

Updated unit test to dynamically verify maxClockRate matches ProductHelper
return value instead of expecting hardcoded 0.

Related-To: NEO-16805

Signed-off-by: Jack Myers <jack.myers@intel.com>
 2025-11-26 01:43:39 +01:00
shubham kumar
1309a77292 feature: Add support for euss buffer overflow on windows 
Related-To: NEO-15684

Signed-off-by: shubham kumar <shubham.kumar@intel.com>
 2025-11-25 20:15:15 +01:00
Mateusz Jablonski
b8e85a9ce4 refactor: move xe3 core definitions to common place 
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
 2025-11-25 16:39:54 +01:00
Compute-Runtime-Validation
074cd41643 Revert "build: update gtest to v1.17.0" 
This reverts commit ebb4f48083.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
 2025-11-25 14:25:11 +01:00
Jaroslaw Warchulski
cc79a136c9 refactor: do not use C headers 
Signed-off-by: Jaroslaw Warchulski <jaroslaw.warchulski@intel.com>
 2025-11-25 12:07:50 +01:00
Szymon Morek
bb15deca7e fix: do not wait on tag when releasing in-place 
Related-To: GSD-11990

If user releases USM without defer policy, do not check whether
allocation is in use when trying to reuse.

Also, when context is being destroyed, use blocking policy to
make sure that tag is updated and driver will notice that
resource is no longer being used by GPU when cleaning up.

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
 2025-11-25 11:12:43 +01:00
Maciej Plewka
c3912acaca fix: do not pass msaa sharing to compile for CL in version below 1.2 
Related-To: NEO-15673
Signed-off-by: Maciej Plewka <maciej.plewka@intel.com>
 2025-11-25 10:35:14 +01:00
Szymon Morek
861ea7200d performance: increase heap size to 4MB on OCL 
Related-To: NEO-16348

Signed-off-by: Szymon Morek <szymon.morek@intel.com>
 2025-11-25 09:58:16 +01:00
Kamil Kopryk
56b30d1803 feature: redesign host function workers 
Each host function gets its unique ID within a CSR,
uses 1 mi store to write ID - to signal that host function is ready,
and 1 mi semaphore wait will wait for the ID to be cleared,
Use 0th bit from ID as pending/completed flag,
host function ID is incremented by 2, and starts with 1.
So each ID will always have 0bit set.
This is a must have since semaphore wait can wait for 4 bytes only.

Adjust command buffer programming and patching logic to IDs.

Add hostFunction callable class - using invoke method,
which stores required information about callback.

Add host function streamer - stores all host function data
for a given CSR.
All user provided host functions are stored in unordered map,
where key is host function ID.

Add host function scheduler, and a thread pool - under debug flag
Single threaded scheduler loops over all registered host function streamers,
dispatch ready to execute host functions to thread pool.

Allow for out of order host functions execution for OOQ - under debug flag,
each host function has bool isInOrder flag which indicates if it can be
executed Out Of Order - in this mode, ID tag will be cleared immediately,
so semaphore wait will unblock before the host function execution.

Remove Host Function worker CV and atomics based implementation.

Rename classes

Related-To: NEO-14577
Signed-off-by: Kamil Kopryk <kamil.kopryk@intel.com>
 2025-11-25 08:05:41 +01:00
Mateusz Jablonski
4f7d2f0315 feature: add support for NVLS 
Related-To: NEO-16649

Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
 2025-11-24 22:33:06 +01:00
Mateusz Jablonski
a22817200f refactor: add wrapper for max gfx core 
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
 2025-11-24 21:35:38 +01:00
Jaroslaw Warchulski
4a8350c29d test: use HWTEST_F instead of HWTEST2_F with MatchAny param 
Signed-off-by: Jaroslaw Warchulski <jaroslaw.warchulski@intel.com>
 2025-11-21 13:17:38 +01:00
Jaroslaw Warchulski
169dad90e1 refactor: add missing #pragma once directives 
Signed-off-by: Jaroslaw Warchulski <jaroslaw.warchulski@intel.com>
 2025-11-21 12:32:55 +01:00
Jaroslaw Warchulski
1ff8a3b4a0 refactor: correct naming of release helper xe2 
Signed-off-by: Jaroslaw Warchulski <jaroslaw.warchulski@intel.com>
 2025-11-21 12:30:05 +01:00
Mateusz Jablonski
ebb4f48083 build: update gtest to v1.17.0 
ignore character-conversion warning in gtest code

Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
 2025-11-21 10:38:16 +01:00
Dominik Dabek
ad767725c5 refactor: move defer internal context init helper 
Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
 2025-11-21 10:02:03 +01:00
Mateusz Jablonski
f33095928c build: rename target gtest to neo_gtest 
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
 2025-11-21 09:41:04 +01:00
Igor Venevtsev
7f593bd295 fix: use condition variables instead of busy waits in worker threads 
Resolves: NEO-16085, GSD-11678, HSD-14025819208

Signed-off-by: Igor Venevtsev <igor.venevtsev@intel.com>
 2025-11-21 08:48:45 +01:00
Bellekallu Rajkiran
2ea3686d86 performance: Make memory resident before cpu access 
Related-To: NEO-13403

Signed-off-by: Bellekallu Rajkiran <bellekallu.rajkiran@intel.com>
 2025-11-21 01:22:31 +01:00
John Falkowski
fbe7a5a887 fix: Disable shared system USM for all KMDless modes 
Related-To: NEO-16531

Signed-off-by: John Falkowski <john.falkowski@intel.com>
 2025-11-20 21:12:02 +01:00
Jaroslaw Warchulski
8c35beb6ef fix: enable state cache invalidation WA for BMG 
Related-To: NEO-16499
Signed-off-by: Jaroslaw Warchulski <jaroslaw.warchulski@intel.com>
 2025-11-20 18:35:00 +01:00
Fabian Zwoliński
4078022318 fix: configure ISA Pool params based on productHelper 
When is2MBLocalMemAlignmentEnabled returns true,
increase pool size for builtins from 64k to 2MB.

Additionally, set appropriate alignment for kernel ISA heap allocations.
Additionally, configure isaAllocationPageSize based on productHelper.

Related-To: NEO-12287
Signed-off-by: Fabian Zwoliński <fabian.zwolinski@intel.com>
 2025-11-20 18:08:14 +01:00
Mateusz Jablonski
553d931f4c build: rename gmock-gtest target to gtest 
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
 2025-11-20 17:32:45 +01:00
Naklicki, Mateusz
8538a0155a feature: update SBA command for SamplerArbitrationControl 
Related-To: NEO-16608
Signed-off-by: Naklicki, Mateusz <mateusz.naklicki@intel.com>
 2025-11-20 17:15:16 +01:00
Bellekallu Rajkiran
864020e12d refactor: Merge duplicate if blocks 
Signed-off-by: Bellekallu Rajkiran <bellekallu.rajkiran@intel.com>
 2025-11-20 16:39:31 +01:00
Jaroslaw Warchulski
cc724285d0 refactor: cleanup includes 
Signed-off-by: Jaroslaw Warchulski <jaroslaw.warchulski@intel.com>
 2025-11-20 16:28:49 +01:00
Jaroslaw Warchulski
f50ca4432d refactor: cleanup includes 
Signed-off-by: Jaroslaw Warchulski <jaroslaw.warchulski@intel.com>
 2025-11-20 08:24:13 +01:00
Mateusz Jablonski
8db92240fc refactor: add wrapper for max product enum value 
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
 2025-11-20 08:04:57 +01:00
Bellekallu Rajkiran
794798d858 performance: Make memory resident before cpu access 
Signed-off-by: Bellekallu Rajkiran <bellekallu.rajkiran@intel.com>
 2025-11-20 07:48:07 +01:00
John Falkowski
45b0886bc4 fix: Disable shared system USM for TBX mode 
Related-To: NEO-16531

Signed-off-by: John Falkowski <john.falkowski@intel.com>
 2025-11-19 20:30:50 +01:00
Neil R. Spruit
b580e4e023 feature: Enable Support for reading updated StateSaveHeaders 
Related-To: NEO-16518

Signed-off-by: Neil R. Spruit <neil.r.spruit@intel.com>
 2025-11-19 16:42:42 +01:00
Jaroslaw Warchulski
a52f526929 refactor: cleanup includes 
Signed-off-by: Jaroslaw Warchulski <jaroslaw.warchulski@intel.com>
 2025-11-19 16:03:24 +01:00
Francois Dugast
703309acda fix: allow execution of clinfo on ARM64 
While this code builds on ARM64, it cannot execute properly because
__rdtsc() is only available on x86. clinfo returns no platform which
gives the incorrect impression that KMD or UMD is not available. A
better fix is still needed but at least clinfo runs.

Signed-off-by: Francois Dugast <francois.dugast@intel.com>
 2025-11-19 15:17:30 +01:00
Filip Hazubski
686672365e fix: Program UseHw64bToken = 1 when initializing WDDM native fence 
Related-To: NEO-16610

Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
 2025-11-19 12:23:47 +01:00
Narendra Bagria
1aefa2b4f6 test: update logger ults 
Related-To: NEO-16500

Signed-off-by: Narendra Bagria <narendra.bagria@intel.com>
 2025-11-19 10:14:32 +01:00
Compute-Runtime-Validation
fe99084f99 Revert "feature: Enable shared system USM by default for BMG" 
This reverts commit d4d9d550b2.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
 2025-11-19 10:02:34 +01:00
Jemale Lockett
dbf9e7ff63 refactor: Add conditions for metadata and vmbind handling 
Resolves: NEO-16089

Signed-off-by: Jemale Lockett <jemale.lockett@intel.com>
 2025-11-18 21:08:04 +01:00
Fabian Zwoliński
1b9b78ac16 feature: implement ISA allocation pooling in OpenCL 
Related-To: NEO-12287
Signed-off-by: Fabian Zwoliński <fabian.zwolinski@intel.com>
 2025-11-18 18:02:51 +01:00
Compute-Runtime-Validation
575f48f738 Revert "refactor: unify stateless and not stateless builtins kernels" 
This reverts commit 4eb37124cb.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
 2025-11-18 16:27:32 +01:00
Kamil Kopryk
c3e98e346a refactor: mark host functions classes as final 
Related-To: NEO-14577
Signed-off-by: Kamil Kopryk <kamil.kopryk@intel.com>
 2025-11-18 14:59:02 +01:00