- Add isaPaddingIncluded flag to AllocationProperties to control ISA
padding
- Modify MemoryManager to skip automatic padding when flag is set
- Update ISAPoolAllocator to manage padding explicitly and set the
flag
- Use actual buffer size for HeapAllocator instead of requested size
- Add test covering different padding scenarios
This prevents double padding addition which was wasting memory when
ISAPoolAllocator managed its own padding on top of MemoryManager's
automatic padding.
Related-To: NEO-12287, HSD-18042276431
Signed-off-by: Fabian Zwoliński <fabian.zwolinski@intel.com>
Move getting usm pool for ptr to its own method.
Move trying to free via pooling to its own method.
Use base ptr of pool for tracking in IPCHandleMap.
Track ipc handle refcount in freeMem.
Return error when trying to use a not allocated ptr within pool.
Related-To: NEO-6893
Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
Allow the application to force storageInfo.localOnly and get the
out-of-memory returned if not possible.
This is a windows-only feature supported on discrete platforms.
Related-To: NEO-13428
Signed-off-by: Maciej Bielski <maciej.bielski@intel.com>
Use device compression defaults when allocating pool.
Ignore compression hints when allocating from pool.
Also remove unused late initialization methods.
Related-To: NEO-6893
Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
Join thread in DllMain (which we are not aware of)
can result in hang occurring in DeferredDeleter, if the library is
freed before FreeLibraryAndExitThread call from within the worker thread,
the thread gets stuck, thus the main thread is stuck on worker->join().
Related-To: NEO-14121
Signed-off-by: Oskar Hubert Weber <oskar.hubert.weber@intel.com>
Related-To: NEO-13715
Implements basic functionality for zeCommandListAppendMemoryAdvise for
system allocator
Signed-off-by: Chandio, Bibrak Qamar <bibrak.qamar.chandio@intel.com>
When trimming old allocations in usm reuse start from largest
allocations.
This will reduce memory usage more quickly once max hold time is hit.
Related-To: NEO-6893, NEO-14429
Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
Real allocation size should be used to properly apply limits and allow
more usm reuse hits.
Related-To: NEO-6893
Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
Save svmData on putting into reuse, instead of searching each time.
Change UNRECOVERABLE_IF to DEBUG_BREAK_IF.
Related-To: NEO-6893
Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
Host usm and device usm for igfx checks system memory usage.
Device usm for dgfx checks local memory usage.
If used memory is above limit threshold:
- no new allocations will be saved for reuse
- cleaner will use shorter hold time of 2 seconds
- cleaner will free all eligible allocations, regardless of async
deleter thread having work
Motivation: in case of gfx memory being full, making resident new
allocations will require evictions which leads to massive slowdown on
enqueue calls.
This change aims to minimize cases where extra memory usage from usm
reuse mechanism leads to above situation.
Related-To: NEO-6893, NEO-14160
Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
Cleaner thread will run every 15ms instead of 2s.
Allocations will be held for at least 10s.
If deferred deleter has elements to release, will skip cleaning cache.
Will clean only 1 allocation per cache, per cleaning run.
Related-To: NEO-6893
Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
mechanism for freeing allocations saved for reuse that have not been
used in a given time
Related-To: NEO-13425
Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
