Commit Graph

5014 Commits

Author SHA1 Message Date
Kacper Nowak
999ec9b2ca refactor: Unify logic for getting atomic FP caps 1/n
- Separate logic for fp16/32/46 caps.
- Add aggregated constexprs for local & global caps of given type
- Pass arguments by reference
- Add hwInfo as argument for future refactors
- Add static_asserts in L0 to ensure there is no mismatch between
internal/external caps
Signed-off-by: Kacper Nowak <kacper.nowak@intel.com>
2024-02-07 11:39:36 +01:00
Compute-Runtime-Validation
f6ca565323 Revert "feature: enable use of waitpkg intrinsic"
This reverts commit 2d15517fbe.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-02-06 16:30:26 +01:00
Dominik Dabek
b4a2e243dc performance: accept AOT kernels for indirects
Use indirects detection mechanism for AOT kernels if the detection
version is at least 2.

Related-To: NEO-7712

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2024-02-06 15:59:42 +01:00
Naklicki, Mateusz
eb0b0c2c89 refactor: add missing xe logs
Signed-off-by: Naklicki, Mateusz <mateusz.naklicki@intel.com>
2024-02-06 13:03:46 +01:00
Mateusz Jablonski
fb9d225495 feature: enable dummy blit WA for PVC
cleanup redundant implementation for DG2

Related-To: NEO-9996
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-02-06 10:17:12 +01:00
Kacper Nowak
ab1ffea592 refactor: Create function to check if OCL2.1 caps are enabled or enforced
Signed-off-by: Kacper Nowak <kacper.nowak@intel.com>
2024-02-06 08:27:09 +01:00
Compute-Runtime-Validation
9b649c1582 Revert "feature: enable bindless mode in L0 on dg2"
This reverts commit fbb8c5a98d.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-02-06 01:41:03 +01:00
Dunajski, Bartosz
f31fafb1e2 refactor: improve debug flag to override bcs mocs
Signed-off-by: Dunajski, Bartosz <bartosz.dunajski@intel.com>
2024-02-05 20:23:54 +01:00
Fabian Zwolinski
fbb8c5a98d feature: enable bindless mode in L0 on dg2
Related-To: NEO-7063
Signed-off-by: Fabian Zwolinski <fabian.zwolinski@intel.com>
2024-02-05 09:48:16 +01:00
Kamil Kopryk
a4f7dda98f refactor: Add xe print debug key
Signed-off-by: Kamil Kopryk <kamil.kopryk@intel.com>
2024-02-02 16:39:51 +01:00
Katarzyna Cencelewska
e6ba9766bd feature: add debug flags to force pat index
for cached recouces: OverridePatIndexForCachedTypes
for uncached resouces: OverridePatIndexForUncachedTypes

Related-To: NEO-10157

Signed-off-by: Katarzyna Cencelewska <katarzyna.cencelewska@intel.com>
2024-02-02 16:11:34 +01:00
Brandon Yates
27c089d60d feature: Register ELF for xe debugger
Related-to:  NEO-9674

Signed-off-by: Brandon Yates <brandon.yates@intel.com>
2024-02-02 09:19:19 +01:00
Jitendra Sharma
00b1f1c5b5 fix: set runalone mode in xe only for render and compute
Runalone mode in XE is supported only for RENDER and COMPUTE.

Related-To: NEO-9139

Signed-off-by: Jitendra Sharma <jitendra.sharma@intel.com>
2024-02-02 09:13:17 +01:00
Dominik Dabek
2cad595a0d performance: debug flag for usm host alloc recycle
set ExperimentalEnableHostAllocationCache=1 to recycle host usm
allocations

Related-To: GSD-7497

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2024-02-01 16:47:59 +01:00
Mateusz Jablonski
2237961deb performance: add macro for ULT-only unrecoverable if
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-02-01 15:09:31 +01:00
Francois Dugast
e311ba5597 refactor: Move ownership of engine type to caller of setDefaultEngine
Related-To: GSD-7097

Signed-off-by: Francois Dugast <francois.dugast@intel.com>
2024-02-01 14:20:20 +01:00
Compute-Runtime-Validation
e7b7eb06e4 Revert "fix: store the correct pagesize in SvmAllocationData"
This reverts commit a104d9199d.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-02-01 05:00:45 +01:00
Katarzyna Cencelewska
8ca9141019 fix: remove wa with override pat index on xe_lpg
Related-To: NEO-9657
Signed-off-by: Katarzyna Cencelewska <katarzyna.cencelewska@intel.com>
2024-01-31 16:30:12 +01:00
Katarzyna Cencelewska
d0b009901c fix: use proper gmm resource type for uncache resources
when new coherency model

Resolves: NEO-9657
Signed-off-by: Katarzyna Cencelewska <katarzyna.cencelewska@intel.com>
2024-01-31 15:26:11 +01:00
Filip Hazubski
d920753ca6 fix: Disable related logic when EnableHostAllocationMemPolicy is not set
Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
2024-01-31 13:45:52 +01:00
Kamil Kopryk
6d3a53fe7f refactor: Improve scratch programming in heapless mode
Related-To: NEO-7621
Signed-off-by: Kamil Kopryk <kamil.kopryk@intel.com>
2024-01-31 10:39:39 +01:00
Lu, Wenbin
a104d9199d fix: store the correct pagesize in SvmAllocationData
Related-To: GSD-7103, NEO-9812

Signed-off-by: Lu, Wenbin <wenbin.lu@intel.com>
2024-01-31 09:12:20 +01:00
Compute-Runtime-Validation
fb46066abc Revert "fix: enable cache env variables for level-zero"
This reverts commit 743904d2df.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-01-31 08:33:05 +01:00
Yoon, Young Jin
cbe35d70a5 fix: initialize libnuma only when flag is set
Modified in memory_info.cpp to initialize libnuma only when
EnableHostAllocationMemPolicy is set.

Related-To: NEO-8276
Signed-off-by: Yoon, Young Jin <young.jin.yoon@intel.com>
2024-01-30 18:27:43 +01:00
Maciej Plewka
564e0f0319 performance: Align host mem to 2MB when range is not limited
Signed-off-by: Maciej Plewka <maciej.plewka@intel.com>

Related-To: NEO-10217
2024-01-30 14:43:13 +01:00
Francois Dugast
278ced35dc fix: Use capability table to determine engine type for defaultEngine
Related-To: GSD-7097

Signed-off-by: Francois Dugast <francois.dugast@intel.com>
2024-01-30 14:28:09 +01:00
Fabian Zwolinski
743904d2df fix: enable cache env variables for level-zero
Related-To: NEO-10045
Signed-off-by: Fabian Zwolinski <fabian.zwolinski@intel.com>
2024-01-30 14:01:42 +01:00
Maciej Plewka
620ad5fa89 fix: change freed chunk size when ptr is aligned
Signed-off-by: Maciej Plewka <maciej.plewka@intel.com>

Related-To: NEO-9945
2024-01-30 12:22:28 +01:00
Compute-Runtime-Validation
fa9c79fb63 Revert "refactor: Add GDI profiling"
This reverts commit 524ae7713a.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-01-30 10:47:34 +01:00
Dominik Dabek
dcab4863d5 performance(ocl): calculate max buffer pool count
Set max buffer pool count to use at most 2 percent of device total memory.

Related-To: NEO-9690

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2024-01-30 07:04:36 +01:00
Mateusz Jablonski
da16dad344 fix: don't limit vm bind support based on platform
Related-To: GSD-7097
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
2024-01-29 19:43:15 +01:00
Lukasz Jobczyk
69c9cacd3e feature: Enable ULLS on wddm 2.3
Related-To: NEO-4843

Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2024-01-29 17:33:28 +01:00
Lukasz Jobczyk
58eaece6c9 fix: Make DirectSubmissionInsertExtraMiMemFenceCommands able to force fence
Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2024-01-29 16:43:24 +01:00
Fabian Zwolinski
c51b656d2c fix: request instruction cache invalidation on module destroy
Invalidation is requested on both linux and windows,
on Csr's that used Isa allocation.

Related-To: NEO-10045
Signed-off-by: Fabian Zwolinski <fabian.zwolinski@intel.com>
2024-01-29 16:00:34 +01:00
Lukasz Jobczyk
64175d8bb6 feature: Add ULLS support for WDDM 2.3
Resolves: NEO-4843

Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2024-01-29 15:45:43 +01:00
Lukasz Jobczyk
524ae7713a refactor: Add GDI profiling
Resolves: NEO-9236
Related-To: NEO-10036

Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2024-01-29 11:36:04 +01:00
Maciej Plewka
7728123907 fix: Do not use 2mb alignment for host ptr allocs
Signed-off-by: Maciej Plewka <maciej.plewka@intel.com>

Related-To: NEO-9945
2024-01-29 11:01:00 +01:00
Lukasz Jobczyk
4704cd49a3 fix: Remove fence handling when reuse cmd buffer
Resolves: NEO-10163
Related-To: NEO-7116

Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2024-01-29 09:08:19 +01:00
Compute-Runtime-Validation
63a5b64c7e Revert "fix: Remove fence handling when reuse cmd buffer"
This reverts commit f3bbd70a58.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-01-27 08:28:57 +01:00
Katarzyna Cencelewska
46fb73026e fix: enable direct submission on release 1274
Related-To: NEO-8187
Signed-off-by: Katarzyna Cencelewska <katarzyna.cencelewska@intel.com>
2024-01-26 16:00:03 +01:00
Katarzyna Cencelewska
eec01e500a fix: non-coherency issue on arl
Resolves: HSD-15015200338
Signed-off-by: Katarzyna Cencelewska <katarzyna.cencelewska@intel.com>
2024-01-26 10:26:33 +01:00
Lukasz Jobczyk
f3bbd70a58 fix: Remove fence handling when reuse cmd buffer
Resolves: NEO-10163
Related-To: NEO-7116

Signed-off-by: Lukasz Jobczyk <lukasz.jobczyk@intel.com>
2024-01-26 10:06:24 +01:00
Compute-Runtime-Validation
7d6845f2aa Revert "performance: Remove tlb flush on DG2 Linux"
This reverts commit 9215f3acb5.

Signed-off-by: Compute-Runtime-Validation <compute-runtime-validation@intel.com>
2024-01-26 07:13:22 +01:00
Katarzyna Cencelewska
b4d2170a6d fix: add missing gmm type to getUsageTypeString
Related-To: HSD-18031172224
Signed-off-by: Katarzyna Cencelewska <katarzyna.cencelewska@intel.com>
2024-01-25 14:12:31 +01:00
Katarzyna Cencelewska
67b0b18be3 fix: osAgnostic path for allocate with alignment
Resolves: NEO-9334
Signed-off-by: Katarzyna Cencelewska <katarzyna.cencelewska@intel.com>
2024-01-25 11:03:34 +01:00
Kamil Kopryk
fdb5a53f85 fix: fix scratch programming in heapless mode l0
Related-To: NEO-10107
Signed-off-by: Kamil Kopryk <kamil.kopryk@intel.com>
2024-01-25 10:50:37 +01:00
Brandon Yates
9636c2f0d5 fix: Avoid double free when freeing sip kernels
Reset sipKernels unique_ptr when freeing

Related-to: NEO-8402

Signed-off-by: Brandon Yates <brandon.yates@intel.com>
2024-01-25 08:45:45 +01:00
Zbigniew Zdanowicz
a25eedb5ac feature: add print of cpu flags and address size upon detection
Related-To: NEO-9737

Signed-off-by: Zbigniew Zdanowicz <zbigniew.zdanowicz@intel.com>
2024-01-24 11:03:30 +01:00
Dominik Dabek
6e434e0424 performance(ocl): increase buffer pool size
increase pool size to 2MB and threshold to 1MB
add limit to the number of pools, set to 2

Related-To: NEO-9690

Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
2024-01-24 10:55:29 +01:00
Zbigniew Zdanowicz
2d15517fbe feature: enable use of waitpkg intrinsic
Related-To: NEO-9737

Signed-off-by: Zbigniew Zdanowicz <zbigniew.zdanowicz@intel.com>
2024-01-24 10:46:28 +01:00