[MLIR] Add atan and atan2 lowerings to CUDA intrinsics

[akuegel]

Thanks :)

mlir/lib/Conversion/GPUToNVVM/LowerGpuOpsToNVVMOps.cpp

patterns
                                           NVVM::GridDimYOp, NVVM::GridDimZOp>,
               GPUShuffleOpLowering, GPUReturnOpLowering,
               // Explicitly drop memory space when lowering private memory
               // attributions since NVVM models it as `alloca`s in the default
               // memory space and does not support `alloca`s with addrspace(5).
               GPUFuncOpLowering<0>>(converter);
   patterns.insert<OpToFuncCallLowering<AbsFOp>>(converter, "__nv_fabsf",
                                                 "__nv_fabs");
+  patterns.insert<OpToFuncCallLowering<AtanOp>>(converter, "__nv_atanf",
+                                                "__nv_atan");
+  patterns.insert<OpToFuncCallLowering<Atan2Op>>(converter, "__nv_atan2f",
+                                                 "__nv_atan2");
   patterns.insert<OpToFuncCallLowering<CeilFOp>>(converter, "__nv_ceilf",
                                                  "__nv_ceil");
   patterns.insert<OpToFuncCallLowering<CosOp>>(converter, "__nv_cosf",
                                                "__nv_cos");
   patterns.insert<OpToFuncCallLowering<ExpOp>>(converter, "__nv_expf",
                                                "__nv_exp");
   patterns.insert<OpToFuncCallLowering<FloorFOp>>(converter, "__nv_floorf",
                                                   "__nv_floor");

mlir/test/Conversion/GPUToNVVM/gpu-to-nvvm.mlir

func @gpu_sqrt(%arg_f16 : f16, %arg_f32 : f32, %arg_f64 : f64)
     %result64 = std.sqrt %arg_f64 : f64
     // CHECK: llvm.call @__nv_sqrt(%{{.*}}) : (!llvm.double) -> !llvm.double
     std.return %result16, %result32, %result64 : f16, f32, f64
   }
 }
 
 // -----
 
+gpu.module @test_module {
+  // CHECK: llvm.func @__nv_atanf(!llvm.float) -> !llvm.float
+  // CHECK: llvm.func @__nv_atan(!llvm.double) -> !llvm.double
+  // CHECK-LABEL: func @gpu_atan
+  func @gpu_atan(%arg_f16 : f16, %arg_f32 : f32, %arg_f64 : f64)
+      -> (f16, f32, f64) {
+    %result16 = std.atan %arg_f16 : f16
+    // CHECK: llvm.fpext %{{.*}} : !llvm.half to !llvm.float
+    // CHECK-NEXT: llvm.call @__nv_atanf(%{{.*}}) : (!llvm.float) -> !llvm.float
+    // CHECK-NEXT: llvm.fptrunc %{{.*}} : !llvm.float to !llvm.half
+    %result32 = std.atan %arg_f32 : f32
+    // CHECK: llvm.call @__nv_atanf(%{{.*}}) : (!llvm.float) -> !llvm.float
+    %result64 = std.atan %arg_f64 : f64
+    // CHECK: llvm.call @__nv_atan(%{{.*}}) : (!llvm.double) -> !llvm.double
+    std.return %result16, %result32, %result64 : f16, f32, f64
+  }
+}
+
+// -----
+
+gpu.module @test_module {
+  // CHECK: llvm.func @__nv_atan2f(!llvm.float, !llvm.float) -> !llvm.float
+  // CHECK: llvm.func @__nv_atan2(!llvm.double, !llvm.double) -> !llvm.double
+  // CHECK-LABEL: func @gpu_atan2
+  func @gpu_atan2(%arg_f16 : f16, %arg_f32 : f32, %arg_f64 : f64)
+      -> (f16, f32, f64) {
+    %result16 = std.atan2 %arg_f16, %arg_f16 : f16
+    // CHECK: llvm.fpext %{{.*}} : !llvm.half to !llvm.float
+    // CHECK: llvm.fpext %{{.*}} : !llvm.half to !llvm.float
+    // CHECK-NEXT: llvm.call @__nv_atan2f(%{{.*}}) : (!llvm.float, !llvm.float) -> !llvm.float
+    // CHECK-NEXT: llvm.fptrunc %{{.*}} : !llvm.float to !llvm.half
+    %result32 = std.atan2 %arg_f32, %arg_f32 : f32
+    // CHECK: llvm.call @__nv_atan2f(%{{.*}}) : (!llvm.float, !llvm.float) -> !llvm.float
+    %result64 = std.atan2 %arg_f64, %arg_f64 : f64
+    // CHECK: llvm.call @__nv_atan2(%{{.*}}) : (!llvm.double, !llvm.double) -> !llvm.double
+    std.return %result16, %result32, %result64 : f16, f32, f64
+  }
+}
+
+// -----
+
 // Test that we handled properly operation with SymbolTable other than module op
 gpu.module @test_module {
   "test.symbol_scope"() ({
   // CHECK: test.symbol_scope
   // CHECK: llvm.func @__nv_expf(!llvm.float) -> !llvm.float
   // CHECK: llvm.func @__nv_exp(!llvm.double) -> !llvm.double
   // CHECK-LABEL: func @gpu_exp
     func @gpu_exp(%arg_f32 : f32, %arg_f64 : f64) -> (f32, f64) {