Sync changes with core dev branch

include/aidge/backend/cuda.hpp

@@ -14,6 +14,5 @@
 
 #include "aidge/backend/cuda/data/TensorImpl.hpp"
 #include "aidge/backend/cuda/operator/ConvImpl.hpp"
-#include "aidge/backend/cuda/operator/ProducerImpl.hpp"
 
 #endif /* AIDGE_BACKEND_CUDA_IMPORTS_H_ */
\ No newline at end of file

include/aidge/backend/cuda/operator/ProducerImpl.hpp

-/********************************************************************************
- * Copyright (c) 2023 CEA-List
- *
- * This program and the accompanying materials are made available under the
- * terms of the Eclipse Public License 2.0 which is available at
- * http://www.eclipse.org/legal/epl-2.0.
- *
- * SPDX-License-Identifier: EPL-2.0
- *
- ********************************************************************************/
-
-#ifndef AIDGE_CUDA_OPERATOR_PRODUCERIMPL_H_
-#define AIDGE_CUDA_OPERATOR_PRODUCERIMPL_H_
-
-#include <memory>
-
-#include "aidge/backend/OperatorImpl.hpp"
-#include "aidge/operator/Producer.hpp"
-#include "aidge/utils/Registrar.hpp"
-#include "aidge/utils/Types.h"
-
-namespace Aidge {
-class ProducerImpl_cuda : public OperatorImpl {
-public:
-    ProducerImpl_cuda(const Producer_Op &op) : OperatorImpl(op) {}
-
-    static std::unique_ptr<ProducerImpl_cuda> create(const Producer_Op &op) {
-        return std::make_unique<ProducerImpl_cuda>(op);
-    }
-
-    NbElts_t getNbProducedData(const IOIndex_t outputIdx) const override final;
-    void forward() override;
-};
-
-namespace {
-static Registrar<Producer_Op> registrarProducerImpl_cuda("cuda", Aidge::ProducerImpl_cuda::create);
-}  // namespace
-}  // namespace Aidge
-
-#endif /* AIDGE_CUDA_OPERATOR_PRODUCERIMPL_H_ */

include/aidge/backend/cuda/utils/CudaContext.hpp

@@ -2,8 +2,8 @@
 #define AIDGE_BACKEND_CUDA_CUDA_CONTEXT_H
 
 #include <vector>
-#include <cstdio>
 
+#include "aidge/utils/ErrorHandling.hpp"
 #include "aidge/backend/cuda/utils/CudaUtils.hpp"
 
 namespace Aidge {
@@ -87,7 +87,7 @@ public:
 
         if (cublas_h[dev] == NULL) {
             CHECK_CUBLAS_STATUS(cublasCreate(&cublas_h[dev]));
-            printf("CUBLAS initialized on device #%d\n", dev);
+            fmt::print("CUBLAS initialized on device #{}\n", dev);
         }
 
         return cublas_h[dev];
@@ -113,7 +113,7 @@ public:
 
         if (cudnn_h[dev] == NULL) {
             CHECK_CUDNN_STATUS(cudnnCreate(&cudnn_h[dev]));
-            printf("CUDNN initialized on device #%d\n", dev);
+            fmt::print("CUDNN initialized on device #{}\n", dev);
         }
 
         return cudnn_h[dev];

include/aidge/backend/cuda/utils/CudaUtils.hpp

@@ -11,6 +11,8 @@
 #include <cuda.h>
 #include <cudnn.h>
 
+#include "aidge/utils/ErrorHandling.hpp"
+
 #define CHECK_CUDNN_STATUS(status)                                             \
     do {                                                                       \
         const cudnnStatus_t e = (status);                                      \

src/operator/ProducerImpl.cpp

-/********************************************************************************
- * Copyright (c) 2023 CEA-List
- *
- * This program and the accompanying materials are made available under the
- * terms of the Eclipse Public License 2.0 which is available at
- * http://www.eclipse.org/legal/epl-2.0.
- *
- * SPDX-License-Identifier: EPL-2.0
- *
- ********************************************************************************/
-
-#include <cassert>
-#include <numeric> // std::accumulate
-#include <vector>
-
-#include "aidge/data/Tensor.hpp"
-#include "aidge/operator/Producer.hpp"
-#include "aidge/utils/Types.h"
-
-#include "aidge/backend/cuda/operator/ProducerImpl.hpp"
-
-Aidge::DimSize_t Aidge::ProducerImpl_cuda::getNbProducedData(
-    Aidge::IOIndex_t outputIdx) const
-{
-    // Requires the whole tensors, regardless of available data on inputs
-    assert(outputIdx == 0 && "operator has only one output");
-    (void) outputIdx;
-
-    return std::static_pointer_cast<Tensor>(mOp.getRawOutput(0))->size();
-}
-
-void Aidge::ProducerImpl_cuda::forward()
-{
-}

src/utils/CudaUtils.cpp

@@ -40,7 +40,7 @@ void Aidge::Cuda::setMultiDevicePeerAccess(unsigned int size, unsigned int* devi
                     CHECK_CUDA_STATUS(cudaSetDevice(devices[j]));
                     const cudaError_t status = cudaDeviceEnablePeerAccess(devices[i], 0);
                     if (status == cudaErrorPeerAccessAlreadyEnabled) {
-                        printf("Peer access already enabled between device %d and device %d\n", devices[j], devices[i]);
+                        fmt::print("Peer access already enabled between device {} and device {}\n", devices[j], devices[i]);
                     } else {
                         CHECK_CUDA_STATUS(status);
                     }

unit_tests/Test_CastMove.cpp

@@ -19,7 +19,7 @@
 #include "aidge/graph/GraphView.hpp"
 #include "aidge/graph/OpArgs.hpp"
 #include "aidge/scheduler/Scheduler.hpp"
-#include "aidge/recipies/Recipies.hpp"
+#include "aidge/recipes/Recipes.hpp"
 
 #include "aidge/backend/cuda.hpp"