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>
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>
Added getInvertedCompatibilityMapping() utility in AOT namespace
to enable efficient reverse lookup of product compatibility. This change
reduces redundant searches and improves performance for compatibility
fallback queries.
Related-To: NEO-14300,GSD-10568
Signed-off-by: Aleksandra Nizio <aleksandra.nizio@intel.com>
When blocking free of chunk is requested, wait for pool allocation
completion.
In pool cleanup, make freeing pool non-blocking.
Related-To: NEO-6893
Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
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>
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>
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>
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>
