intel/compute-runtime
1
0
Fork 0
mirror of https://github.com/intel/compute-runtime.git synced 2026-01-05 09:09:04 +08:00
Code Issues Packages Projects Releases Wiki Activity

Limit overestimation in multi kernel scenarios.

Browse Source 
- There was overestimation that resulted in each kernel getting
page aligned estimation size.
- After this change every kernel aligns only to cache line and final
size is aligned to page size.

Change-Id: Iee06bdd0083724ea7e9415f3d0fe70198acca407
Signed-off-by: Mrozek, Michal <michal.mrozek@intel.com>
This commit is contained in:
Mrozek, Michal
2019-04-23 15:57:23 +02:00
committed by sys_ocldev
parent 6cf2dc411d
commit c269bc062f
2 changed files with 31 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
runtime/helpers/kernel_commands.inl
View File

@@ -90,9 +90,10 @@ size_t getSizeRequired(const MultiDispatchInfo &multiDispatchInfo, SizeGetterT &
size_t totalSize = 0;
auto it = multiDispatchInfo.begin();
for (auto e = multiDispatchInfo.end(); it != e; ++it) {
totalSize = alignUp(totalSize, MemoryConstants::pageSize);
totalSize = alignUp(totalSize, MemoryConstants::cacheLineSize);
totalSize += getSize(*it, std::forward<ArgsT>(args)...);
}
totalSize = alignUp(totalSize, MemoryConstants::pageSize);
return totalSize;
}