AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

P4 to Git Change 1793462 by gandryey@gera-win10 on 2019/06/07 15:33:19

Browse Source 
SWDEV-79445 - OCL generic changes and code clean-up
	Optimize scratch buffer calculation in the preparation for coop group launch, since the current limit affects max waves calculation:
	- Switch to 32 waves per CU as the max possible limit
	- Use vgprs count for the waves limit calculation to avoid unconditional possible max

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/pal/paldevice.cpp#141 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/paldevice.hpp#38 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/pal/palsettings.cpp#82 edit
This commit is contained in:
foreman
2019-06-07 16:04:41 -04:00
parent 180c320d5c
commit f829f695e3
3 changed files with 30 additions and 22 deletions
Download Patch File Download Diff File Expand all files Collapse all files
rocclr/runtime/device/pal/palsettings.cpp
+1 -1
View File
@@ -112,7 +112,7 @@ Settings::Settings() {
numDeviceEvents_ = 1024;
numWaitEvents_ = 8;
numScratchWavesPerCu_ = 16;
numScratchWavesPerCu_ = 32;
// Don't support platform atomics by default.
svmAtomics_ = false;