P4 to Git Change 1313578 by lmoriche@lmoriche_opencl_dev on 2016/09/13 00:14:36

SWDEV-94610 - Target features are only needed in the CL->IR stage. The attributes remain on the function, so they should not be set again in the IR->ISA stage.

Affected files ...

... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/roccompiler.cpp#16 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/rocm/rocprogram.cpp#34 edit


[ROCm/clr commit: a1009a5d11]

projects/clr/rocclr/runtime/device/rocm/roccompiler.cpp

@@ -196,8 +196,16 @@ HSAILProgram::compileImpl_LC(
         return false;
     }
 
+    // Set fp32-denormals and fp64-denormals
+    bool fp32Denormals = !options->oVariables->DenormsAreZero
+        && dev().deviceInfo().gfxipVersion_ >= 900;
+
+    driverOptions.append(" -Xclang -target-feature -Xclang ");
+    driverOptions.append(fp32Denormals ? "+" : "-")
+        .append("fp32-denormals,+fp64-denormals");
+
     // Tokenize the options string into a vector of strings
-    std::istringstream istrstr(driverOptions + codegenOptions(options));
+    std::istringstream istrstr(driverOptions);
     std::istream_iterator<std::string> sit(istrstr), end;
     std::vector<std::string> params(sit, end);
 

projects/clr/rocclr/runtime/device/rocm/rocprogram.cpp

@@ -914,7 +914,7 @@ HSAILProgram::linkImpl_LC(amd::option::Options *options)
     codegenOptions.append(" ").append(optLevel.str());
 
     // Tokenize the options string into a vector of strings
-    std::istringstream strstr(codegenOptions + this->codegenOptions(options));
+    std::istringstream strstr(codegenOptions);
     std::istream_iterator<std::string> sit(strstr), end;
     std::vector<std::string> params(sit, end);
 
@@ -1438,14 +1438,7 @@ HSAILProgram::codegenOptions(amd::option::Options* options)
 {
     std::string optionsStr;
 
-#if defined(WITH_LIGHTNING_COMPILER)
-    bool fp32Denormals = !options->oVariables->DenormsAreZero
-        && dev().deviceInfo().gfxipVersion_ >= 900;
-
-    optionsStr.append(" -Xclang -target-feature -Xclang ");
-    optionsStr.append(fp32Denormals ? "+" : "-")
-        .append("fp32-denormals,+fp64-denormals");
-#else // !defined(WITH_LIGHTNING_COMPILER)
+#if !defined(WITH_LIGHTNING_COMPILER)
     if (dev().deviceInfo().gfxipVersion_ < 900) {
         optionsStr.append(" -cl-denorms-are-zero");
     }