P4 to Git Change 1159349 by rili@rili_opencl_stg on 2015/06/09 12:52:32

EPR #419313 - Blit Kernel execution split
	                         Enabled flag GPU_ENABLE_LARGE_ALLOCATION
	                         Return error if global work size is larger than 32bits.

	Code Review#7664

Affected files ...

... //depot/stg/opencl/drivers/opencl/api/opencl/amdocl/cl_execute.cpp#22 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpukernel.cpp#285 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpukernel.hpp#111 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuvirtual.cpp#365 edit
... //depot/stg/opencl/drivers/opencl/runtime/utils/flags.hpp#234 edit


[ROCm/clr commit: ca7f6ab119]

projects/clr/rocclr/runtime/device/gpu/gpukernel.cpp

@@ -3561,6 +3561,7 @@ HSAILKernel::init(bool finalize)
     //compile kernel down to ISA
     if (finalize) {
         std::string options(compileOptions_.c_str());
+        flags_.internalKernel_ = (compileOptions_.find("-cl-internal-kernel") != 0) ? true: false;
         options.append(" -just-kernel=");
         options.append(openClKernelName.c_str());
         // Append an option so that we can selectively enable a SCOption on CZ