Extend initial support for primitive types in PTX backend

- Allow i16, i32, i64, float, and double types, using the native .u16,
  .u32, .u64, .f32, and .f64 PTX types.
- Allow loading/storing of all primitive types.
- Allow primitive types to be passed as parameters.
- Allow selection of PTX Version and Shader Model as sub-target attributes.
- Merge integer/floating-point test cases for load/store.
- Use .u32 instead of .s32 to conform to output from NVidia nvcc compiler.

Patch by Justin Holewinski



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@126824 91177308-0d34-0410-b5e6-96231b3b80d8

test/CodeGen/PTX/mul.ll

@@ -10,16 +10,30 @@
 ;	ret i32 %z
 ;}
 
-define ptx_device float @t3(float %x, float %y) {
+define ptx_device float @t1_f32(float %x, float %y) {
 ; CHECK: mul.f32 f0, f1, f2
 ; CHECK-NEXT: ret;
   %z = fmul float %x, %y
   ret float %z
 }
 
-define ptx_device float @t4(float %x) {
+define ptx_device double @t1_f64(double %x, double %y) {
+; CHECK: mul.f64 fd0, fd1, fd2
+; CHECK-NEXT: ret;
+  %z = fmul double %x, %y
+  ret double %z
+}
+
+define ptx_device float @t2_f32(float %x) {
 ; CHECK: mul.f32 f0, f1, 0F40A00000;
 ; CHECK-NEXT: ret;
   %z = fmul float %x, 5.0
   ret float %z
 }
+
+define ptx_device double @t2_f64(double %x) {
+; CHECK: mul.f64 fd0, fd1, 0D4014000000000000;
+; CHECK-NEXT: ret;
+  %z = fmul double %x, 5.0
+  ret double %z
+}