From 3b45e88e4c6a61ea482a7ee4634fed28979800a6 Mon Sep 17 00:00:00 2001
From: Zonglin Peng <zonglinpeng@fb.com>
Date: Mon, 3 Mar 2025 16:08:50 -0800
Subject: [PATCH 1/2] init

---
 backends/cadence/aot/functions.yaml | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/backends/cadence/aot/functions.yaml b/backends/cadence/aot/functions.yaml
index f1a5b6a50b0..173538cc2ee 100644
--- a/backends/cadence/aot/functions.yaml
+++ b/backends/cadence/aot/functions.yaml
@@ -214,6 +214,11 @@
     - arg_meta: null
       kernel_name: impl::reference::quantized_relu_out
 
+- func: cadence::quantized_relu.per_tensor_out(Tensor X, int X_zero_point, int out_zero_point, int out_multiplier, int out_shift, *, Tensor(a!) out) -> Tensor(a!)
+  kernels:
+    - arg_meta: null
+      kernel_name: impl::reference::quantized_relu_per_tensor_out
+
 - func: cadence::quantized_matmul.out(Tensor X, int X_zero_point, Tensor Y, int Y_zero_point, Tensor? bias, int out_multiplier, int out_shift, int out_zero_point, bool transposed, *, Tensor(a!) out) -> Tensor(a!)
   kernels:
     - arg_meta: null

From c29492079a12b83235a5f41ad0f901831e0dc831 Mon Sep 17 00:00:00 2001
From: Zonglin Peng <zonglinpeng@fb.com>
Date: Mon, 3 Mar 2025 16:09:33 -0800
Subject: [PATCH 2/2] init

---
 .../operators/quantized_relu_out.cpp          | 54 +++++++++++++++++++
 1 file changed, 54 insertions(+)

diff --git a/backends/cadence/reference/operators/quantized_relu_out.cpp b/backends/cadence/reference/operators/quantized_relu_out.cpp
index 19b971405c9..7a385849aee 100644
--- a/backends/cadence/reference/operators/quantized_relu_out.cpp
+++ b/backends/cadence/reference/operators/quantized_relu_out.cpp
@@ -7,6 +7,7 @@
  */
 
 #include <executorch/backends/cadence/reference/kernels/kernels.h>
+#include <executorch/backends/cadence/reference/operators/operators.h>
 #include <executorch/runtime/kernel/kernel_includes.h>
 
 namespace impl {
@@ -75,6 +76,59 @@ void quantized_relu_out(
   }
 }
 
+template <typename T>
+void quantized_relu_per_tensor_out_(
+    __ET_UNUSED KernelRuntimeContext& ctx,
+    const Tensor& input,
+    const int64_t in_zero_point,
+    const int64_t out_zero_point,
+    const int64_t out_multiplier,
+    const int64_t out_shift,
+    Tensor& output) {
+  const T* __restrict__ in = input.const_data_ptr<T>();
+  T* __restrict__ out = output.mutable_data_ptr<T>();
+
+  // Compute the out_scale from out_multiplier and out_shift
+  const float out_scale = -out_multiplier * 1.0 / (1 << 31) * pow(2, out_shift);
+
+  for (size_t i = 0, e = input.numel(); i < e; ++i) {
+    const float temp = in[i] > in_zero_point ? (in[i] - in_zero_point) : 0;
+    out[i] = kernels::quantize<T>(temp, out_scale, out_zero_point);
+  }
+}
+
+void quantized_relu_per_tensor_out(
+    KernelRuntimeContext& ctx,
+    const Tensor& input,
+    const int64_t in_zero_point,
+    const int64_t out_zero_point,
+    const int64_t out_multiplier,
+    const int64_t out_shift,
+    Tensor& output) {
+#define typed_quantized_relu(ctype, dtype)    \
+  case executorch::aten::ScalarType::dtype: { \
+    quantized_relu_per_tensor_out_<ctype>(    \
+        ctx,                                  \
+        input,                                \
+        in_zero_point,                        \
+        out_zero_point,                       \
+        out_multiplier,                       \
+        out_shift,                            \
+        output);                              \
+    break;                                    \
+  }
+
+  executorch::aten::ScalarType dtype = input.scalar_type();
+  switch (dtype) {
+    ET_FORALL_CADENCE_QUANTIZED_TYPES(typed_quantized_relu)
+    default:
+      ET_DCHECK_MSG(
+          false, "Unhandled dtype %s", torch::executor::toString(dtype));
+  }
+
+#undef typed_quantized_relu
+}
+
 }; // namespace native
 }; // namespace reference
 }; // namespace impl