[EE/BE] add float variant to unary_ufunc_realhbbf16_to_floathbf16 (#12289)

This PR was created by the merge bot to help merge the original PR into
the main branch.
ghstack PR number: #12277 by
@manuelcandales
^ Please use this as the source of truth for the PR details, comments,
and reviews
ghstack PR base:
https://github.com/pytorch/executorch/tree/gh/manuelcandales/130/base
ghstack PR head:
https://github.com/pytorch/executorch/tree/gh/manuelcandales/130/head
Merge bot PR base: https://github.com/pytorch/executorch/tree/main
Merge bot PR head:
https://github.com/pytorch/executorch/tree/gh/manuelcandales/130/orig
@diff-train-skip-merge

Co-authored-by: Manuel Candales <[email protected]>

Author: pytorchbot

backends/cadence/fusion_g3/operators/op_exp.cpp

@@ -60,7 +60,7 @@ Tensor& exp_out(KernelRuntimeContext& ctx, const Tensor& in, Tensor& out) {
     return out;
   } else {
     return torch::executor::native::internal::
-        unary_ufunc_realhbbf16_to_floathbf16(std::exp, ctx, in, out);
+        unary_ufunc_realhbbf16_to_floathbf16(std::exp, std::exp, ctx, in, out);
   }
 }
 

backends/cadence/fusion_g3/operators/op_rsqrt.cpp

@@ -27,7 +27,8 @@ namespace native {
 
 namespace {
 
-double rsqrt(double x) {
+template <typename T>
+T rsqrt(T x) {
   return 1.0 / std::sqrt(x);
 }
 
@@ -61,11 +62,11 @@ Tensor& rsqrt_out(KernelRuntimeContext& ctx, const Tensor& in, Tensor& out) {
     return out;
   } else {
     return torch::executor::native::internal::
-        unary_ufunc_realhbbf16_to_floathbf16(rsqrt, ctx, in, out);
+        unary_ufunc_realhbbf16_to_floathbf16(rsqrt, rsqrt, ctx, in, out);
   }
 }
 
 } // namespace native
 } // namespace G3
 } // namespace impl
-} // namespace cadence
+} // namespace cadence

backends/cadence/fusion_g3/operators/op_sqrt.cpp

@@ -55,7 +55,8 @@ Tensor& sqrt_out(KernelRuntimeContext& ctx, const Tensor& in, Tensor& out) {
     return out;
   } else {
     return torch::executor::native::internal::
-        unary_ufunc_realhbbf16_to_floathbf16(std::sqrt, ctx, in, out);
+        unary_ufunc_realhbbf16_to_floathbf16(
+            std::sqrt, std::sqrt, ctx, in, out);
   }
 }
 

backends/cadence/fusion_g3/operators/op_tanh.cpp

@@ -55,7 +55,8 @@ Tensor& tanh_out(KernelRuntimeContext& ctx, const Tensor& in, Tensor& out) {
     return out;
   } else {
     return torch::executor::native::internal::
-        unary_ufunc_realhbbf16_to_floathbf16(std::tanh, ctx, in, out);
+        unary_ufunc_realhbbf16_to_floathbf16(
+            std::tanh, std::tanh, ctx, in, out);
   }
 }
 

backends/cadence/hifi/operators/op_rsqrt.cpp

@@ -21,7 +21,8 @@ namespace HiFi {
 namespace native {
 namespace {
 
-double rsqrt(double x) {
+template <typename T>
+T rsqrt(T x) {
   return 1.0 / std::sqrt(x);
 }
 
@@ -46,7 +47,7 @@ Tensor& rsqrt_out(RuntimeContext& ctx, const Tensor& in, Tensor& out) {
   }
 
   return torch::executor::native::internal::
-      unary_ufunc_realhbbf16_to_floathbf16(rsqrt, ctx, in, out);
+      unary_ufunc_realhbbf16_to_floathbf16(rsqrt, rsqrt, ctx, in, out);
 }
 
 } // namespace native

backends/cadence/hifi/operators/op_tanh.cpp

@@ -35,10 +35,10 @@ Tensor& tanh_out(RuntimeContext& ctx, const Tensor& in, Tensor& out) {
   }
 
   return torch::executor::native::internal::
-      unary_ufunc_realhbbf16_to_floathbf16(std::tanh, ctx, in, out);
+      unary_ufunc_realhbbf16_to_floathbf16(std::tanh, std::tanh, ctx, in, out);
 }
 
 } // namespace native
 } // namespace HiFi
 } // namespace impl
-} // namespace cadence
+} // namespace cadence

kernels/portable/cpu/op_acos.cpp

@@ -14,10 +14,7 @@ namespace torch {
 namespace executor {
 namespace native {
 
-Tensor& acos_out(KernelRuntimeContext& ctx, const Tensor& in, Tensor& out) {
-  return internal::unary_ufunc_realhbbf16_to_floathbf16(
-      std::acos, ctx, in, out);
-}
+DEFINE_UNARY_UFUNC_REALHBBF16_TO_FLOATHBF16(acos_out, std::acos)
 
 } // namespace native
 } // namespace executor

kernels/portable/cpu/op_acosh.cpp

@@ -14,10 +14,7 @@ namespace torch {
 namespace executor {
 namespace native {
 
-Tensor& acosh_out(KernelRuntimeContext& ctx, const Tensor& in, Tensor& out) {
-  return internal::unary_ufunc_realhbbf16_to_floathbf16(
-      std::acosh, ctx, in, out);
-}
+DEFINE_UNARY_UFUNC_REALHBBF16_TO_FLOATHBF16(acosh_out, std::acosh)
 
 } // namespace native
 } // namespace executor

kernels/portable/cpu/op_asin.cpp

@@ -14,10 +14,7 @@ namespace torch {
 namespace executor {
 namespace native {
 
-Tensor& asin_out(KernelRuntimeContext& ctx, const Tensor& in, Tensor& out) {
-  return internal::unary_ufunc_realhbbf16_to_floathbf16(
-      std::asin, ctx, in, out);
-}
+DEFINE_UNARY_UFUNC_REALHBBF16_TO_FLOATHBF16(asin_out, std::asin)
 
 } // namespace native
 } // namespace executor

kernels/portable/cpu/op_asinh.cpp

@@ -14,10 +14,7 @@ namespace torch {
 namespace executor {
 namespace native {
 
-Tensor& asinh_out(KernelRuntimeContext& ctx, const Tensor& in, Tensor& out) {
-  return internal::unary_ufunc_realhbbf16_to_floathbf16(
-      std::asinh, ctx, in, out);
-}
+DEFINE_UNARY_UFUNC_REALHBBF16_TO_FLOATHBF16(asinh_out, std::asinh)
 
 } // namespace native
 } // namespace executor