Instead of storing AIL configurations in global table, store it
in root device environment.
This also prevents potential scenario with accessing deleted memory due
to symbol collision when application uses both OCL/L0 libraries.
- AIL is now stored in root device environment, and gets initialized
with other helpers
- Minor: corrected naming in ULTs
Signed-off-by: Kacper Nowak <kacper.nowak@intel.com>
Related-To: NEO-9240
Refactored eviction mechanism works as follows:
- eviction is needed only if
total size of cache binaries + size of the new binary exceed cache limit
- single evition call removes files with a summed size of 1/3 of the cache limit
- if new binary can not fit in the cache size limit
even after eviction, it will not be saved
- cache limit applies only to
files in cache directory with .cl_cache/.l0_cache extension.
Only these files are counted and only these files are removed
Minor:
- rename variables for better readability
- add `const` where possible
Related-To: NEO-4262
Signed-off-by: Fabian Zwolinski <fabian.zwolinski@intel.com>
add attribute to override cacheable attribute to gmm constructor
enable this override for command buffers on mtl
change command buffers back to allocation by kmd
this keeps the quicker allocation which is needed to keep enqueue times
low
Related-To: NEO-8152
Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
Instead of storing AIL configurations in global table, store it
in root device environment.
This also prevents potential scenario with accessing deleted memory due
to symbol collision when application uses both OCL/L0 libraries.
- AIL is now stored in root device environment, and gets initialized
with other helpers
- Minor: corrected naming in ULTs
Signed-off-by: Kacper Nowak <kacper.nowak@intel.com>
Related-To: NEO-9240
when direct submission and monitor fence is not required then use
current fence value to update flush stamp
Related-To: NEO-9123
Signed-off-by: Katarzyna Cencelewska <katarzyna.cencelewska@intel.com>
- surfaceStateSize is in pages, bindless size needs to be programmed in
surface state units
Related-To: NEO-7063
Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
- usingSurfaceStateHeap indicates if any of the args is using local ssh
in bindless kernels:
without global allocator - ssh is used for all args
with global bindless allocator - ssh used only for buffer with offset
set in surface state, otherwise not used
When any of the args is using ssh - getSurfaceStateHeapDataSize() returns
non-zero size.
Related-To: NEO-7063
Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
Device allocation chunking only applies for multi-tile mode for implicit scaling
Related-To: NEO-9051
Signed-off-by: John Falkowski <john.falkowski@intel.com>
Added an extension version of zeContextGetVirtualAddressSpaceIpcHandle
and zeContextPutVirtualAddressSpaceIpcHandle, which return
ZE_RESULT_ERROR_UNSUPPORTED_FEATURE for now.
Related-To: NEO-8259
Signed-off-by: Young Jin Yoon <young.jin.yoon@intel.com>
Signed-off-by: Michal Mrozek <michal.mrozek@intel.com>
Related-To: NEO-6989
-Prevent imbalance in multi dimensional dispatches
-Make sure to utilize as much Eus as possible
-Prefer highest possible tg dspatch count possible
-Make sure that xe_core doesn't have uneven workgroups
Added an extension version of zeContextGetVirtualAddressSpaceIpcHandle
and zeContextPutVirtualAddressSpaceIpcHandle, which return
ZE_RESULT_ERROR_UNSUPPORTED_FEATURE for now.
Related-To: NEO-8259
Signed-off-by: Young Jin Yoon <young.jin.yoon@intel.com>
Compute-Runtime-Validation