intel/compute-runtime
1
0
Fork 0
mirror of https://github.com/intel/compute-runtime.git synced 2025-12-24 21:18:24 +08:00
Code Issues Packages Projects Releases Wiki Activity

XeHPC Implicit scaling: put command/ring/semaphore buffer to first memory bank

Browse Source 
In direct submission scenario command/ring/semaphore buffer allocations
are placed in the same memory bank to ensure that their memory is updated in
correct order

Related-To: NEO-6698
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
This commit is contained in:
Mateusz Jablonski
2022-03-16 18:41:29 +00:00
committed by Compute-Runtime-Automation
parent aa7ba69746
commit 3792481d33
7 changed files with 337 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
shared/source/direct_submission/direct_submission_hw.inl
View File

@@ -66,7 +66,7 @@ bool DirectSubmissionHw<GfxFamily, Dispatcher>::allocateResources() {
const AllocationProperties commandStreamAllocationProperties{device.getRootDeviceIndex(),
true, allocationSize,
AllocationType::RING_BUFFER,
isMultiOsContextCapable, osContext.getDeviceBitfield()};
isMultiOsContextCapable, false, osContext.getDeviceBitfield()};
ringBuffer = memoryManager->allocateGraphicsMemoryWithProperties(commandStreamAllocationProperties);
UNRECOVERABLE_IF(ringBuffer == nullptr);
allocations.push_back(ringBuffer);
@@ -78,7 +78,7 @@ bool DirectSubmissionHw<GfxFamily, Dispatcher>::allocateResources() {
const AllocationProperties semaphoreAllocationProperties{device.getRootDeviceIndex(),
true, MemoryConstants::pageSize,
AllocationType::SEMAPHORE_BUFFER,
isMultiOsContextCapable, osContext.getDeviceBitfield()};
isMultiOsContextCapable, false, osContext.getDeviceBitfield()};
semaphores = memoryManager->allocateGraphicsMemoryWithProperties(semaphoreAllocationProperties);
UNRECOVERABLE_IF(semaphores == nullptr);
allocations.push_back(semaphores);