From 9eadee681c5f6ba16ab68657582fb3878e82ab37 Mon Sep 17 00:00:00 2001
From: Olivier BICHLER <olivier.bichler@cea.fr>
Date: Thu, 6 Feb 2025 15:44:32 +0100
Subject: [PATCH] Use double instead of float to process integers
---
.../cpu/operator/GlobalAveragePoolingImpl_kernels.hpp | 6 +++---
.../aidge/backend/cpu/operator/ReduceMeanImpl_kernels.hpp | 6 +++---
2 files changed, 6 insertions(+), 6 deletions(-)
diff --git a/include/aidge/backend/cpu/operator/GlobalAveragePoolingImpl_kernels.hpp b/include/aidge/backend/cpu/operator/GlobalAveragePoolingImpl_kernels.hpp
index 4a95c3a4..d5e5561d 100644
--- a/include/aidge/backend/cpu/operator/GlobalAveragePoolingImpl_kernels.hpp
+++ b/include/aidge/backend/cpu/operator/GlobalAveragePoolingImpl_kernels.hpp
@@ -40,9 +40,9 @@ stableMean(const T* vec, size_t size) {
template <typename T>
typename std::enable_if<!std::is_floating_point<T>::value, T>::type
stableMean(const T* vec, size_t size) {
- float mean = 0;
+ double mean = 0;
for (size_t i = 0; i < size; ++i) {
- mean = std::fma<float>(vec[i] - mean, 1.0f / (i + 1), mean);
+ mean = std::fma<double>(vec[i] - mean, 1.0f / (i + 1), mean);
}
return mean;
}
@@ -55,7 +55,7 @@ castFromFloat(T value) {
template <typename T>
typename std::enable_if<!std::is_floating_point<T>::value, T>::type
-castFromFloat(float value) {
+castFromFloat(double value) {
return static_cast<T>(std::nearbyint(value));
}
diff --git a/include/aidge/backend/cpu/operator/ReduceMeanImpl_kernels.hpp b/include/aidge/backend/cpu/operator/ReduceMeanImpl_kernels.hpp
index c0962408..864b89c4 100644
--- a/include/aidge/backend/cpu/operator/ReduceMeanImpl_kernels.hpp
+++ b/include/aidge/backend/cpu/operator/ReduceMeanImpl_kernels.hpp
@@ -40,9 +40,9 @@ stableMean(const T* vec, size_t len, size_t stride) {
template <typename T>
typename std::enable_if<!std::is_floating_point<T>::value, T>::type
stableMean(const T* vec, size_t len, size_t stride) {
- float mean = 0;
+ double mean = 0;
for (size_t i = 0; i < len; ++i) {
- mean = std::fma<float>(vec[i * stride] - mean, 1.0f / (i + 1), mean);
+ mean = std::fma<double>(vec[i * stride] - mean, 1.0f / (i + 1), mean);
}
return mean;
}
@@ -55,7 +55,7 @@ castFromFloat(T value) {
template <typename T>
typename std::enable_if<!std::is_floating_point<T>::value, T>::type
-castFromFloat(float value) {
+castFromFloat(double value) {
return static_cast<T>(std::nearbyint(value));
}
--
GitLab