Making TorchForce CUDA-graph aware

README.md

@@ -252,6 +252,43 @@ to return forces.
 torch_force.setOutputsForces(True)
 ```
 
+Recording the model into a CUDA graph
+-------------------------------------
+
+You can ask `TorchForce` to run the model using [CUDA graphs](https://pytorch.org/docs/stable/notes/cuda.html#cuda-graphs). Not every model will be compatible with this feature, but it can be a significant performance boost for some models. To enable it the CUDA platform must be used and an special property must be provided to `TorchForce`:
+
+```python
+torch_force.setProperty("useCUDAGraphs", "true")
+# The property can also be set at construction
+torch_force = TorchForce('model.pt', {'useCUDAGraphs': 'true'})
+```
+
+The first time the model is run, it will be compiled (also known as recording) into a CUDA graph. Subsequent runs will use the compiled graph, which can be significantly faster. It is possible that compilation fails, in which case an `OpenMMException` will be raised. If that happens, you can disable CUDA graphs and try again.
+
+It is required to run the model at least once before recording, in what is known as warmup.
+By default ```TorchForce``` will run the model just once before recording, but longer warming up might be desired. In these cases one can set the property ```CUDAGraphWarmupSteps```:
+```python
+torch_force.setProperty("CUDAGraphWarmupSteps", "12")
+```
+
+List of available properties
+----------------------------
+
+Some ```TorchForce``` functionalities can be customized by setting properties on an instance of it. Properties can be set at construction or by using ```setProperty```. A property is a pair of key/value strings. For instance:
+
+```python
+torch_force = TorchForce('model.pt', {'useCUDAGraphs': 'true'})
+torch_force.setProperty("useCUDAGraphs", "true")
+print("Current properties:")
+for property in torch_force.getProperties():
+    print(property.key, property.value)
+```
+
+Currently, the following properties are available:
+
+1. useCUDAGraphs: Turns on the CUDA graph functionality
+2. CUDAGraphWarmupSteps: When CUDA graphs are being used, controls the number of warmup calls to the model before recording.
+
 License
 =======
 

openmmapi/include/TorchForce.h

@@ -11,7 +11,7 @@
  *                                                                            *
  * Portions copyright (c) 2018-2022 Stanford University and the Authors.      *
  * Authors: Peter Eastman                                                     *
- * Contributors:                                                              *
+ * Contributors: Raimondas Galvelis, Raul P. Pelaez                           *
  *                                                                            *
  * Permission is hereby granted, free of charge, to any person obtaining a    *
  * copy of this software and associated documentation files (the "Software"), *
@@ -34,6 +34,7 @@
 
 #include "openmm/Context.h"
 #include "openmm/Force.h"
+#include <map>
 #include <string>
 #include <torch/torch.h>
 #include "internal/windowsExportTorch.h"
@@ -54,17 +55,20 @@ class OPENMM_EXPORT_NN TorchForce : public OpenMM::Force {
      * Create a TorchForce.  The network is defined by a PyTorch ScriptModule saved
      * to a file.
      *
-     * @param file   the path to the file containing the network
+     * @param file       the path to the file containing the network
+     * @param properties optional list of properties
      */
-    TorchForce(const std::string& file);
+    TorchForce(const std::string& file,
+               const std::map<std::string, std::string>& properties = {});
     /**
      * Create a TorchForce.  The network is defined by a PyTorch ScriptModule
      * Note that this constructor makes a copy of the provided module.
      * Any changes to the module  after calling this constructor will be ignored by TorchForce.
      *
      * @param module   an instance of the torch module
+     * @param properties optional list of properties
      */
-    TorchForce(const torch::jit::Module &module);
+    TorchForce(const torch::jit::Module &module, const std::map<std::string, std::string>& properties = {});
     /**
      * Get the path to the file containing the network.
      * If the TorchForce instance was constructed with a module, instead of a filename,
@@ -140,6 +144,19 @@ class OPENMM_EXPORT_NN TorchForce : public OpenMM::Force {
      * @param defaultValue   the default value of the parameter
      */
     void setGlobalParameterDefaultValue(int index, double defaultValue);
+    /**
+     * Set a value of a property.
+     *
+     * @param name           the name of the property
+     * @param value          the value of the property
+     */
+    void setProperty(const std::string& name, const std::string& value);
+    /**
+     * Get the list of properties for this instance.
+     * An unset property has an empty string as its value.
+     * @return A map of property names to values.
+     */
+    const std::map<std::string, std::string>& getProperties() const;
 protected:
     OpenMM::ForceImpl* createImpl() const;
 private:
@@ -148,6 +165,8 @@ class OPENMM_EXPORT_NN TorchForce : public OpenMM::Force {
     bool usePeriodic, outputsForces;
     std::vector<GlobalParameterInfo> globalParameters;
     torch::jit::Module module;
+    std::map<std::string, std::string> properties;
+    std::string emptyProperty;
 };
 
 /**

openmmapi/src/TorchForce.cpp

@@ -8,7 +8,7 @@
  *                                                                            *
  * Portions copyright (c) 2018-2022 Stanford University and the Authors.      *
  * Authors: Peter Eastman                                                     *
- * Contributors:                                                              *
+ * Contributors: Raimondas Galvelis, Raul P. Pelaez                           *
  *                                                                            *
  * Permission is hereby granted, free of charge, to any person obtaining a    *
  * copy of this software and associated documentation files (the "Software"), *
@@ -41,10 +41,17 @@ using namespace TorchPlugin;
 using namespace OpenMM;
 using namespace std;
 
-TorchForce::TorchForce(const torch::jit::Module& module) : file(), usePeriodic(false), outputsForces(false), module(module) {
+TorchForce::TorchForce(const torch::jit::Module& module, const map<string, string>& properties) : file(), usePeriodic(false), outputsForces(false), module(module) {
+    const std::map<std::string, std::string> defaultProperties = {{"useCUDAGraphs", "false"}, {"CUDAGraphWarmupSteps", "1"}};
+    this->properties = defaultProperties;
+    for (auto& property : properties) {
+        if (defaultProperties.find(property.first) == defaultProperties.end())
+            throw OpenMMException("TorchForce: Unknown property '" + property.first + "'");
+        this->properties[property.first] = property.second;
+    }
 }
 
-TorchForce::TorchForce(const std::string& file) : TorchForce(torch::jit::load(file)) {
+TorchForce::TorchForce(const std::string& file, const map<string, string>& properties) : TorchForce(torch::jit::load(file), properties) {
     this->file = file;
 }
 
@@ -78,7 +85,7 @@ bool TorchForce::getOutputsForces() const {
 
 int TorchForce::addGlobalParameter(const string& name, double defaultValue) {
     globalParameters.push_back(GlobalParameterInfo(name, defaultValue));
-    return globalParameters.size()-1;
+    return globalParameters.size() - 1;
 }
 
 int TorchForce::getNumGlobalParameters() const {
@@ -104,3 +111,13 @@ void TorchForce::setGlobalParameterDefaultValue(int index, double defaultValue)
     ASSERT_VALID_INDEX(index, globalParameters);
     globalParameters[index].defaultValue = defaultValue;
 }
+
+void TorchForce::setProperty(const std::string& name, const std::string& value) {
+    if (properties.find(name) == properties.end())
+        throw OpenMMException("TorchForce: Unknown property '" + name + "'");
+    properties[name] = value;
+}
+
+const std::map<std::string, std::string>& TorchForce::getProperties() const {
+    return properties;
+}