Commit Graph

3241 Commits

Author SHA1 Message Date
10be59cb15 Improve isIpSamplingSupported helper
Signed-off-by: Kamil Kopryk <kamil.kopryk@intel.com>
Related-To: NEO-6738
2022-04-20 17:02:20 +02:00
b81380a879 Enable engines round robin assign
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2022-04-20 11:52:33 +02:00
93fe8a4794 Optimize gpgpu csr's mutex lock in the enqueue blit
Signed-off-by: Jobczyk, Lukasz <lukasz.jobczyk@intel.com>
2022-04-20 11:40:59 +02:00
7f08a9d993 Remove adjustHwInfoForIgc helper
Signed-off-by: Kamil Kopryk <kamil.kopryk@intel.com>
Related-To: NEO-6738
2022-04-20 10:04:36 +02:00
b5b8762faf refactor direct submission: Pass command stream receiver to ctor
Related-To: NEO-6643

Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2022-04-19 23:02:55 +02:00
e47385dc11 Refactor ShaderChannelSelect logic for images using CL_LUMINANCE
Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2022-04-19 15:58:38 +02:00
6fdcb83020 Add missing tbx calls to timestamp wait logic
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2022-04-19 14:24:06 +02:00
1a33fc0dc7 Zebin: return error on parsing empty vector entries in zeInfo
This commit prevents a yaml parsing error in case a data type is passed
after empty vetor type data entry with the same indendation. In this
case, a parsing error will be returned.
- Corrected .ze_info section in valid empty program (zebin mock)
- Minor ults refactor in order to use mock zebin program with valid
.ze_info

Related-To: NEO-6735
Signed-off-by: Kacper Nowak <kacper.nowak@intel.com>
2022-04-19 13:12:47 +02:00
15fe7e06ce Add reporting of extension cl_intel_bfloat16_conversions
Resolves: NEO-6885

Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2022-04-19 12:33:08 +02:00
a8680c0b82 Implement ULTs for IgaWrapper
This change:
- implements mock for functions from IGA DLL
- implements ULTs for IgaWrapper
- implements RAII wrapper for GTEST's capture of stdout
- adds ASAN and TSAN flags to ocloc_tests

Related-To: NEO-6834
Signed-off-by: Patryk Wrobel <patryk.wrobel@intel.com>
2022-04-19 12:26:36 +02:00
303c12bc5b Refactor and test initialization of IGC in ocloc
This change:
- encapsulates initialization of IGC in a separate class
- removes code duplication from OfflineLinker and OfflineCompiler
- tests the new encapsulated class

Related-To: NEO-6834
Signed-off-by: Patryk Wrobel <patryk.wrobel@intel.com>
2022-04-19 12:15:34 +02:00
5f37aa4ba4 Pass cache settings to aub_stream
Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2022-04-19 11:00:07 +02:00
f6f5baf23a Revert "Fix Scratch surface state programming on PVC"
This reverts commit 5b175ebb4a.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2022-04-19 10:53:22 +02:00
34ebcc9520 Zebin: Use addend in relocations
feat(zebin): Use addend from RELA sections when performing relocations.

Resolves: NEO-6898

Signed-off-by: Krystian Chmielewski <krystian.chmielewski@intel.com>
2022-04-19 09:19:26 +02:00
4bebb477d4 Improve coverage of ocloc_api
This change contains ULTs, which test the code
which was not covered.

Related-To: NEO-6834
Signed-off-by: Patryk Wrobel <patryk.wrobel@intel.com>
2022-04-15 16:11:08 +02:00
7e4262839a Add timestamp wait support for BCS
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2022-04-15 13:26:23 +02:00
c9d61840a2 Allocate SBA buffers per HW context
- different physical storage for every HW context
- adds support for debugging with implicit scaling on
- reorganize tests

Relates-To: NEO-6883

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2022-04-15 10:01:28 +02:00
8d1ad5a4f3 Refactor: use stack vector for root device indices
Stack vector will not cause dynamic allocations in most circumstances
ie. number of root device indices not more than 16

Related-To: NEO-6837

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2022-04-14 14:05:42 +02:00
352583b9d9 Detect GPU hang in evictUnusedAllocations()
This change introduces checking of the return value
of wait function in case of blocking version of
evictUnusedAllocations(). Furthermore, it propagates
the error to the callers. It contains also ULTs.

Related-To: NEO-6681
Signed-off-by: Patryk Wrobel <patryk.wrobel@intel.com>
2022-04-13 15:39:02 +02:00
a54c947866 test: create dedicated mock object for timestamp packet
Related-To: NEO-6871

Signed-off-by: Zbigniew Zdanowicz <zbigniew.zdanowicz@intel.com>
2022-04-13 14:56:36 +02:00
ab69e87a5e Add report of extension cl_intel_split_work_group_barrier
Related-To: NEO-6746

Signed-off-by: Wesierski, Lukasz <lukasz.wesierski@intel.com>
2022-04-13 14:37:37 +02:00
3e19f39b8d Add RPL-P device IDs
Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2022-04-13 09:59:58 +02:00
7d4d22228f Check for null allocation in processResources
Signed-off-by: Jaime Arteaga <jaime.a.arteaga.molina@intel.com>
2022-04-13 02:36:27 +02:00
858766c080 Remove unused members and getters in system info
Related-To: NEO-6835

Signed-off-by: Slawomir Milczarek <slawomir.milczarek@intel.com>
2022-04-12 10:46:53 +02:00
789587bacc Return nullptr for gfx alloc mmap error
Signed-off-by: John Falkowski <john.falkowski@intel.com>
2022-04-12 10:08:19 +02:00
ee0d183cf9 Handle legacy hasBarriers properly
Previous change regarding NEO-6785 added encoding of number of barriers
to specific value representation depending on hardware that we program for.

In patch token format encoding of number of barriers is sent via
hasBarriers field in a token.
In zebin true number of barriers is sent via barrier_count field in
zeInfo.

To remove this discrepancy, translate encoded number of barriers into
true number of barriers in legacy format.

Resolves: NEO-6785

Signed-off-by: Krystian Chmielewski <krystian.chmielewski@intel.com>
2022-04-12 09:44:10 +02:00
884d729e4e Improve pat index programming on linux
Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2022-04-12 08:18:20 +02:00
01a719cf33 Allow for zebin rebuild when IR is present
This commits adds rebuilding zebin binary.
If zebin is built for different device and has SPIRV, then new ze binary
will be built using SPIRV.

Signed-off-by: Krystian Chmielewski <krystian.chmielewski@intel.com>
2022-04-12 08:05:25 +02:00
e9ebd10ab8 Revert "Correct ShaderChannelSelect value for images using CL_LUMINANCE"
This reverts commit 6f9ce94a17.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2022-04-12 06:47:24 +02:00
6f9ce94a17 Correct ShaderChannelSelect value for images using CL_LUMINANCE
Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2022-04-11 17:13:57 +02:00
ce645f13b7 Encode PRODUCT_CONFIG value into fatbinary
Change modifies the encoding entry in fatbinary for platforms.
If numbering in -device is used, the value PRODUCT_CONFIG will be encoded.
The functionality that returns the correct product config values has
also been added.

Related-To: NEO-6744
Signed-off-by: Daria Hinz <daria.hinz@intel.com>
2022-04-11 15:09:17 +02:00
9e7703a578 Fix Wddm DirectSubmission unit tests
Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2022-04-11 12:42:58 +02:00
6f911ba840 Add unit tests for MultiCommand class
This change improves code coverage of MultiCommand class.

Related-To: NEO-6834
Signed-off-by: Patryk Wrobel <patryk.wrobel@intel.com>
2022-04-08 22:05:16 +02:00
667594e8eb Revert "Improve multi storage handling."
This reverts commit 81739c0265.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2022-04-08 13:07:27 +02:00
2c1bfbb5b2 Encode number barriers
When programming number of barriers use BARRIER_SIZE enumeration.
Resolves: NEO-6785

Signed-off-by: Krystian Chmielewski <krystian.chmielewski@intel.com>
2022-04-08 10:32:23 +02:00
81739c0265 Improve multi storage handling.
- color always basing on granularity
- align all chunks to granularity size
- make all chunks equal in size

Signed-off-by: Michal Mrozek <michal.mrozek@intel.com>
2022-04-08 10:21:34 +02:00
4d4ccfd128 Remove cache info base class
There is only one implementation of said class and we don't
even adhere to the interface it provides.

Signed-off-by: Daniel Chabrowski <daniel.chabrowski@intel.com>
2022-04-07 18:15:57 +02:00
fc4eaa1894 Refactor passing GlobalFenceAllocation to DirectSubmission
Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2022-04-07 16:08:07 +02:00
76349b8030 Improve code coverage of OfflineCompiler class
This change introduces unit tests related to
member functions of OfflineCompiler class.
OfflineCompiler::initialize() is not covered
and it will be added in a separate commit.

Related-To: NEO-6834
Signed-off-by: Patryk Wrobel <patryk.wrobel@intel.com>
2022-04-07 13:58:52 +02:00
db9c0d1103 Refactor and enable MI_MEM_FENCE programming for DirectSubmission dispatch
Signed-off-by: Bartosz Dunajski <bartosz.dunajski@intel.com>
2022-04-07 12:53:56 +02:00
fd45ac133d Unify append calls
Related-To: NEO-6242

Signed-off-by: Zbigniew Zdanowicz <zbigniew.zdanowicz@intel.com>
2022-04-07 10:18:55 +02:00
5b175ebb4a Fix Scratch surface state programming on PVC
- program max num of threads in surface state

Resolves: NEO-6850

Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
2022-04-07 07:44:23 +02:00
7a3976ad64 Revert "Force 64KB page size for cpu alignment in dual storage allocation"
This reverts commit 7ff6a5c1fa.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2022-04-07 00:45:07 +02:00
64d3f95410 Revert "Allow for zebin rebuild when IR is present"
This reverts commit 2ae3f3e521.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2022-04-06 23:17:22 +02:00
7ff6a5c1fa Force 64KB page size for cpu alignment in dual storage allocation
There is no need to force 2MB alignment for CPU allocation in dual
storage usage. Additionaly for WSL this will allow to avoid usage of
malloc in driver path.

Relates-To: NEO-6620
Signed-off-by: Kamil Diedrich <kamil.diedrich@intel.com>
2022-04-06 15:28:17 +02:00
f4407064a4 Refactor store register mem encoder to include partition parameter
Related-To: NEO-6811

Signed-off-by: Zbigniew Zdanowicz <zbigniew.zdanowicz@intel.com>
2022-04-06 14:00:56 +02:00
76289be8c1 Avoid unnecessary allocation in enqueueHandler
Related-To: NEO-6837

kernelObjsForAuxTranslation will allocate only if needed

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2022-04-06 13:03:41 +02:00
9e1bd1005b Disable aub test
Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2022-04-06 09:06:26 +02:00
eaba246464 Test prelim command stream and clos cache
Signed-off-by: Daniel Chabrowski daniel.chabrowski@intel.com
Related-To: NEO-6591
2022-04-05 16:48:00 +02:00
2ae3f3e521 Allow for zebin rebuild when IR is present
Signed-off-by: Krystian Chmielewski <krystian.chmielewski@intel.com>
2022-04-05 10:25:06 +02:00