[ET][Portable] Check tensor dtype inside elementwise_utils

Pull Request resolved: pytorch/executorch#6008


ghstack-source-id: 246985130
@exported-using-ghexport

Differential Revision: [D64052137](https://our.internmc.facebook.com/intern/diff/D64052137/)

Author: manuelcandales

kernels/portable/cpu/op_add.cpp

@@ -22,14 +22,6 @@ Tensor& add_out(
     const Tensor& b,
     const Scalar& alpha,
     Tensor& out) {
-  ET_KERNEL_CHECK(
-      ctx,
-      (executorch::runtime::tensor_is_realhbbf16_type(a) &&
-       executorch::runtime::tensor_is_realhbbf16_type(b) &&
-       executorch::runtime::tensor_is_realhbbf16_type(out)),
-      InvalidArgument,
-      out);
-
   // Common Dtype
   ScalarType common_type = promoteTypes(a.scalar_type(), b.scalar_type());
 
@@ -64,6 +56,7 @@ Tensor& add_out(
         [val_alpha](const CTYPE_COMPUTE val_a, const CTYPE_COMPUTE val_b) {
           return val_a + val_alpha * val_b;
         },
+        ctx,
         a,
         utils::SupportedTensorDtypes::REALHBBF16,
         b,
@@ -81,13 +74,6 @@ Tensor& add_scalar_out(
     const Scalar& b,
     const Scalar& alpha,
     Tensor& out) {
-  ET_KERNEL_CHECK(
-      ctx,
-      (executorch::runtime::tensor_is_realhbbf16_type(a) &&
-       executorch::runtime::tensor_is_realhbbf16_type(out)),
-      InvalidArgument,
-      out);
-
   // Common Dtype
   ScalarType common_type = utils::promote_type_with_scalar(a.scalar_type(), b);
 
@@ -120,6 +106,7 @@ Tensor& add_scalar_out(
           CTYPE_COMPUTE val_alpha = utils::scalar_to<CTYPE_COMPUTE>(alpha);
           return val_a + val_alpha * val_b;
         },
+        ctx,
         a,
         utils::SupportedTensorDtypes::REALHBBF16,
         out,

kernels/portable/cpu/op_clamp.cpp

@@ -73,13 +73,6 @@ Tensor& clamp_out(
     const exec_aten::optional<Scalar>& min_opt,
     const exec_aten::optional<Scalar>& max_opt,
     Tensor& out) {
-  ET_KERNEL_CHECK(
-      ctx,
-      (executorch::runtime::tensor_is_realhbbf16_type(in) &&
-       executorch::runtime::tensor_is_realhbbf16_type(out)),
-      InvalidArgument,
-      out);
-
   bool has_min = min_opt.has_value();
   bool has_max = max_opt.has_value();
 
@@ -154,6 +147,7 @@ Tensor& clamp_out(
           }
           return val_out;
         },
+        ctx,
         in,
         utils::SupportedTensorDtypes::REALHBBF16,
         out,
@@ -182,15 +176,6 @@ Tensor& clamp_tensor_out(
   const Tensor& min = has_min ? min_opt.value() : in;
   const Tensor& max = has_max ? max_opt.value() : in;
 
-  ET_KERNEL_CHECK(
-      ctx,
-      (executorch::runtime::tensor_is_realhbbf16_type(in) &&
-       executorch::runtime::tensor_is_realhbbf16_type(min) &&
-       executorch::runtime::tensor_is_realhbbf16_type(max) &&
-       executorch::runtime::tensor_is_realhbbf16_type(out)),
-      InvalidArgument,
-      out);
-
   // Common Dtype
   ScalarType common_type = in.scalar_type();
   if (has_min) {
@@ -239,6 +224,7 @@ Tensor& clamp_tensor_out(
           }
           return val_out;
         },
+        ctx,
         in,
         utils::SupportedTensorDtypes::REALHBBF16,
         min,

kernels/portable/cpu/op_where.cpp

@@ -19,16 +19,6 @@ Tensor& where_out(
     const Tensor& a,
     const Tensor& b,
     Tensor& out) {
-  ET_KERNEL_CHECK(
-      ctx,
-      ((cond.scalar_type() == ScalarType::Bool ||
-        cond.scalar_type() == ScalarType::Byte) &&
-       executorch::runtime::tensor_is_realhbbf16_type(a) &&
-       executorch::runtime::tensor_is_realhbbf16_type(b) &&
-       executorch::runtime::tensor_is_realhbbf16_type(out)),
-      InvalidArgument,
-      out);
-
   // Common Dtype
   ScalarType common_type = promoteTypes(a.scalar_type(), b.scalar_type());
 
@@ -57,6 +47,7 @@ Tensor& where_out(
         [](const CTYPE_COMPUTE val_a,
            const CTYPE_COMPUTE val_b,
            const CTYPE_COMPUTE val_c) { return val_c ? val_a : val_b; },
+        ctx,
         a,
         utils::SupportedTensorDtypes::REALHBBF16,
         b,

kernels/portable/cpu/util/elementwise_util.cpp

@@ -0,0 +1,49 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under the BSD-style license found in the
+ * LICENSE file in the root directory of this source tree.
+ */
+
+#include <executorch/kernels/portable/cpu/util/elementwise_util.h>
+
+namespace torch {
+namespace executor {
+namespace native {
+namespace utils {
+namespace internal {
+
+bool check_tensor_dtype(
+    const Tensor t,
+    SupportedTensorDtypes dtypes,
+    const ScalarType compute_type) {
+  switch (dtypes) {
+    case SupportedTensorDtypes::REALHBBF16:
+      return executorch::runtime::tensor_is_realhbbf16_type(t);
+    case SupportedTensorDtypes::BOOL_OR_BYTE:
+      return (
+          executorch::runtime::tensor_is_type(t, ScalarType::Bool) ||
+          executorch::runtime::tensor_is_type(t, ScalarType::Byte));
+    case SupportedTensorDtypes::SAME_AS_COMPUTE:
+      return executorch::runtime::tensor_is_type(t, compute_type);
+    case SupportedTensorDtypes::SAME_AS_COMMON: {
+      if (compute_type == ScalarType::Float) {
+        return (
+            executorch::runtime::tensor_is_type(t, ScalarType::Float) ||
+            executorch::runtime::tensor_is_type(t, ScalarType::Half) ||
+            executorch::runtime::tensor_is_type(t, ScalarType::BFloat16));
+      } else {
+        return executorch::runtime::tensor_is_type(t, compute_type);
+      }
+    }
+  }
+  ET_CHECK(false);
+  return false;
+}
+
+} // namespace internal
+} // namespace utils
+} // namespace native
+} // namespace executor
+} // namespace torch

kernels/portable/cpu/util/elementwise_util.h

@@ -229,15 +229,29 @@ store_common_to_tensor_fn<CTYPE_COMMON> get_store_common_to_tensor_fn(
   return nullptr;
 }
 
+bool check_tensor_dtype(
+    const Tensor t,
+    SupportedTensorDtypes dtypes,
+    const ScalarType compute_type);
+
 } // namespace internal
 
 template <typename CTYPE_COMMON, const char* op_name, typename Op>
 inline void apply_unitensor_elementwise_fn(
     const Op& compute_fun,
+    KernelRuntimeContext& ctx,
     const Tensor& a,
     SupportedTensorDtypes a_dtypes,
     const Tensor& out,
     SupportedTensorDtypes out_dtypes) {
+  constexpr auto compute_type = CppTypeToScalarType<CTYPE_COMMON>::value;
+
+  ET_KERNEL_CHECK(
+      ctx,
+      (internal::check_tensor_dtype(a, a_dtypes, compute_type) &&
+       internal::check_tensor_dtype(out, out_dtypes, compute_type)),
+      InvalidArgument, );
+
   const auto load_a_to_common =
       internal::get_load_to_common_fn<CTYPE_COMMON, op_name>(a, a_dtypes);
   const auto store_common_to_out =
@@ -263,12 +277,22 @@ inline void apply_unitensor_elementwise_fn(
 template <typename CTYPE_COMMON, const char* op_name, typename Op>
 inline void apply_bitensor_elementwise_fn(
     const Op& compute_fun,
+    KernelRuntimeContext& ctx,
     const Tensor& a,
     SupportedTensorDtypes a_dtypes,
     const Tensor& b,
     SupportedTensorDtypes b_dtypes,
     const Tensor& out,
     SupportedTensorDtypes out_dtypes) {
+  constexpr auto compute_type = CppTypeToScalarType<CTYPE_COMMON>::value;
+
+  ET_KERNEL_CHECK(
+      ctx,
+      (internal::check_tensor_dtype(a, a_dtypes, compute_type) &&
+       internal::check_tensor_dtype(b, b_dtypes, compute_type) &&
+       internal::check_tensor_dtype(out, out_dtypes, compute_type)),
+      InvalidArgument, );
+
   const bool a_is_broadcasted = !out.sizes().equals(a.sizes());
   const bool b_is_broadcasted = !out.sizes().equals(b.sizes());
   const bool any_is_broadcasted = (a_is_broadcasted || b_is_broadcasted);
@@ -312,9 +336,9 @@ inline void apply_bitensor_elementwise_fn(
 }
 
 /**
- * Useful for tri-tensor elementwise operators. For each element of the inputs,
- * perform a computation and write to the corresponding element of the output.
- * Tensor broadcasting is applied wherever it is required.
+ * Useful for tri-tensor elementwise operators. For each element of the
+ * inputs, perform a computation and write to the corresponding element of the
+ * output. Tensor broadcasting is applied wherever it is required.
  *
  * In order to mitigate build time cost (straightforwardly |CTYPE_A| *
  * |CTYPE_B| * |CTYPE_C| * |CTYPE_OUT|), all arguments to compute_fun
@@ -334,6 +358,7 @@ inline void apply_bitensor_elementwise_fn(
 template <typename CTYPE_COMMON, const char* op_name, typename Op>
 inline void apply_tritensor_elementwise_fn(
     const Op& compute_fun,
+    KernelRuntimeContext& ctx,
     const Tensor& a,
     SupportedTensorDtypes a_dtypes,
     const Tensor& b,
@@ -342,6 +367,16 @@ inline void apply_tritensor_elementwise_fn(
     SupportedTensorDtypes c_dtypes,
     const Tensor& out,
     SupportedTensorDtypes out_dtypes) {
+  constexpr auto compute_type = CppTypeToScalarType<CTYPE_COMMON>::value;
+
+  ET_KERNEL_CHECK(
+      ctx,
+      (internal::check_tensor_dtype(a, a_dtypes, compute_type) &&
+       internal::check_tensor_dtype(b, b_dtypes, compute_type) &&
+       internal::check_tensor_dtype(c, c_dtypes, compute_type) &&
+       internal::check_tensor_dtype(out, out_dtypes, compute_type)),
+      InvalidArgument, );
+
   const bool a_is_broadcasted = !out.sizes().equals(a.sizes());
   const bool b_is_broadcasted = !out.sizes().equals(b.sizes());
   const bool c_is_broadcasted = !out.sizes().equals(c.sizes());

kernels/portable/cpu/util/targets.bzl

@@ -80,6 +80,7 @@ def define_common_targets():
 
     runtime.cxx_library(
         name = "elementwise_util",
+        srcs = ["elementwise_util.cpp"],
         exported_headers = [
             "elementwise_util.h",
         ],

runtime/core/exec_aten/util/tensor_util.h

@@ -469,6 +469,16 @@ inline bool tensor_is_bool_type(exec_aten::Tensor t) {
   return true;
 }
 
+inline bool tensor_is_type(exec_aten::Tensor t, exec_aten::ScalarType dtype) {
+  ET_LOG_MSG_AND_RETURN_IF_FALSE(
+      t.scalar_type() == dtype,
+      "Expected to find %s type, but tensor has type %s",
+      torch::executor::toString(dtype),
+      torch::executor::toString(t.scalar_type()));
+
+  return true;
+}
+
 inline bool tensor_is_integral_type(
     exec_aten::Tensor t,
     bool includeBool = false) {