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

P4 to Git Change 1079952 by yaxunl@yaxunl_stg_win50 on 2014/09/23 12:31:16

Browse Source 
ECR #377625 - Workaround for Blender performance issue. Lower available VGPRs to improve waves per CU.

	Added BuildOptsAppend to OCL app profile.
	Read BuildOptsAppend and append to build options.
	Added specific wave optimization option for Blender.

Affected files ...

... //depot/stg/opencl/drivers/opencl/appprofiles/oclappprofile.xml#7 edit
... //depot/stg/opencl/drivers/opencl/compiler/lib/backends/gpu/scwrapper/SI/scCompileSI.cpp#45 edit
... //depot/stg/opencl/drivers/opencl/compiler/lib/utils/OPTIONS.def#116 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/appprofile.cpp#10 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/appprofile.hpp#8 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/device.cpp#170 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/device.hpp#230 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuappprofile.cpp#10 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuappprofile.hpp#6 edit
... //depot/stg/opencl/drivers/opencl/runtime/platform/program.cpp#63 edit
This commit is contained in:
foreman
2014-09-23 12:44:50 -04:00
parent 5ef908c5e7
commit 16f8ca9aae
8 changed files with 119 additions and 76 deletions
Download Patch File Download Diff File Expand all files Collapse all files
rocclr/runtime/device/gpu/gpuappprofile.cpp
-54
View File
@@ -21,59 +21,5 @@ AppProfile::AppProfile()
PropertyData(DataType_Boolean, &reportAsOCL12Device_)));
}
bool AppProfile::ParseApplicationProfile()
{
amd::ADL* adl = new amd::ADL;
if ((adl == NULL) || !adl->init()) {
delete adl;
return false;
}
ADLApplicationProfile* pProfile = NULL;
// Apply blb configurations
int result = adl->adl2ApplicationProfilesProfileOfApplicationx2Search(
adl->adlContext(), wsAppFileName_.c_str(), NULL, NULL,
L"OCL", &pProfile);
delete adl;
if (pProfile == NULL) {
return false;
}
PropertyRecord* firstProperty = pProfile->record;
uint32_t valueOffset = 0;
for (int index = 0; index < pProfile->iCount; index++) {
PropertyRecord* profileProperty = reinterpret_cast<PropertyRecord*>
((reinterpret_cast<char*>(firstProperty)) + valueOffset);
// Get property name
char* propertyName = profileProperty->strName;
auto entry = propertyDataMap_.find(std::string(propertyName));
if (entry == propertyDataMap_.end()) {
// unexpected name
valueOffset += (sizeof(PropertyRecord) + profileProperty->iDataSize - 4);
continue;
}
// Get the property value
switch (entry->second.type_) {
case DataType_Boolean:
*(reinterpret_cast<bool*>(entry->second.data_)) =
profileProperty->uData[0] ? true : false;
break;
default:
break;
}
valueOffset += (sizeof(PropertyRecord) + profileProperty->iDataSize - 4);
}
free(pProfile);
return true;
}
}
rocclr/runtime/device/gpu/gpuappprofile.hpp
-19
View File
@@ -19,26 +19,7 @@ public:
bool enableHighPerformanceState() const { return enableHighPerformanceState_; }
bool reportAsOCL12Device() const { return reportAsOCL12Device_; }
protected:
//! parse application profile based on application file name
virtual bool ParseApplicationProfile();
private:
enum DataTypes
{
DataType_Unknown = 0,
DataType_Boolean,
};
struct PropertyData {
PropertyData(DataTypes type, void* data): type_(type), data_(data) {}
DataTypes type_; //!< Data type
void* data_; //!< Pointer to the data
};
typedef std::map<std::string, PropertyData> DataMap;
DataMap propertyDataMap_;
bool enableHighPerformanceState_;
bool reportAsOCL12Device_;