refactor: moved cmake_module folder, updated README, shortened test s…

…yntax
mspronesti · Feb 10, 2022 · 59d514d · 59d514d
1 parent 7e21a57
commit 59d514d
Show file tree

Hide file tree

Showing 5 changed files with 40 additions and 36 deletions.
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -10,7 +10,7 @@ project(baylib)
 set(CMAKE_CXX_STANDARD 20)
 set(CMAKE_CXX_STANDARD_REQUIRED ON)
 set(CMAKE_CXX_EXTENSIONS OFF)
-set(CMAKE_MODULE_PATH ${CMAKE_CURRENT_SOURCE_DIR}/baylib/cmake_module)
+set(CMAKE_MODULE_PATH cmake_module)
 
 # change this option to compile the examples
 option(BUILD_EXAMPLES "Build the examples showing how to use baylib" OFF)

diff --git a/README.md b/README.md
@@ -11,7 +11,8 @@ Baylib is a parallel inference library for discrete Bayesian networks supporting
 ## Main features
 Here's a list of the main requested features:
 * Copy-On-Write semantics for the graph data structure, including the conditional probability table (CPT) of each node
-* parallel implementation of the algorithms either using C++17 threads or GPU computing with [boost compute](https://www.boost.org/doc/libs/1_66_0/libs/compute/doc/html/index.html)
+* parallel implementation of the algorithms either using C++17 threads or GPGPUU optimization
+* GPGPU optimization implemented with opencl, using [boost compute](https://www.boost.org/doc/libs/1_66_0/libs/compute/doc/html/index.html) and cuda.
 * template-based classes for probability format
 * input compatibility with the [XDSL format](https://support.bayesfusion.com/docs/) provided by the SMILE library
 * cmake-based deployment
@@ -23,13 +24,13 @@ Here's a list of the main requested features:
 * Rejection Sampling - C++11 threads
 * Adaptive importance sampling - C++11 threads, GPGPU with boost compute
 
-|       algorithm      	         | evidence 	| deterministic nodes 	| multi-threading 	| GPGPU 	|
-|:------------------------------:|--------------|-----------------------|-------------------|-----------|
-| gibbs sampling       	         |    &check;   |     *                 |     &check;       |           |
-| likelihood weighting 	         |    &check;   |     &check;           |     &check;       |           |
-| logic sampling       	         |    &check;   |     &check;           |                   |  &check;  |
-| rejection sampling  	         |    &check;   |     &check;           |     &check;       |           |
-| adaptive importance sampling   |    &check;  	|     &check;           |     &check;       |  &check;  |
+|       algorithm      	         | evidence 	| deterministic nodes 	| multi-threading 	| GPGPU-opencl 	| GPGPU - CUDA |
+|:------------------------------:|--------------|-----------------------|-------------------|---------------|--------------|
+| gibbs sampling       	         |    &check;   |     *                 |     &check;       |               |              |
+| likelihood weighting 	         |    &check;   |     &check;           |     &check;       |               |              |
+| logic sampling       	         |    &check;   |     &check;           |                   |  &check;      | &check;      |
+| rejection sampling  	         |    &check;   |     &check;           |     &check;       |               |              |
+| adaptive importance sampling   |    &check;  	|     &check;           |     &check;       |  &check;      |              |
 
 *It's a very well-known limitation of the Gibbs sampling approach
 ## Dependencies

diff --git a/baylib/cmake_module/FindThrust.cmake → cmake/FindThrust.cmake b/baylib/cmake_module/FindThrust.cmake → cmake/FindThrust.cmake
@@ -26,7 +26,7 @@
 #
 
 find_path( THRUST_INCLUDE_DIR
-        HINTS ./
+        HINTS .
         ../thrust
         ../../thrust
         ../../../thrust

diff --git a/test/evidence_test.cpp b/test/evidence_test.cpp
@@ -22,32 +22,34 @@
 
 using namespace baylib::inference;
 using probability_type = double;
+template<class Variable>
+using bnet = baylib::bayesian_net<Variable>;
 
 
 
 
 template<typename Probability, class Variable>
-        std::vector<baylib::marginal_distribution<Probability>> get_results(const baylib::bayesian_net<Variable> &bn){
+        std::vector<baylib::marginal_distribution<Probability>> get_results(const bnet<Variable> &bn){
             std::vector<baylib::marginal_distribution<Probability>> results{
-                logic_sampling<baylib::bayesian_net<Variable>>(bn, SAMPLES, MEMORY).make_inference(),
-                gibbs_sampling<baylib::bayesian_net<Variable>>(bn, SAMPLES, THREADS).make_inference(),
-                likelihood_weighting<baylib::bayesian_net<Variable>>(bn, SAMPLES, THREADS).make_inference(),
-                rejection_sampling<baylib::bayesian_net<Variable>>(bn, SAMPLES, THREADS).make_inference(),
-                adaptive_importance_sampling<baylib::bayesian_net<Variable>>(bn, SAMPLES, MEMORY).make_inference(),
+                logic_sampling<bnet<Variable>>(bn, SAMPLES, MEMORY).make_inference(),
+                gibbs_sampling<bnet<Variable>>(bn, SAMPLES, THREADS).make_inference(),
+                likelihood_weighting<bnet<Variable>>(bn, SAMPLES, THREADS).make_inference(),
+                rejection_sampling<bnet<Variable>>(bn, SAMPLES, THREADS).make_inference(),
+                adaptive_importance_sampling<bnet<Variable>>(bn, SAMPLES, MEMORY).make_inference(),
 #ifdef CUDA_CMP_FOUND
-                logic_sampling_cuda<baylib::bayesian_net<Variable>>(bn, SAMPLES).make_inference()
+                logic_sampling_cuda<bnet<Variable>>(bn, SAMPLES).make_inference()
 #endif
             };
             return results;
         }
 
 template<typename Probability, class Variable>
-        std::vector<baylib::marginal_distribution<Probability>> get_results_heavy(const baylib::bayesian_net<Variable> &bn){
+        std::vector<baylib::marginal_distribution<Probability>> get_results_heavy(const bnet<Variable> &bn){
             std::vector<baylib::marginal_distribution<Probability>> results{
-                likelihood_weighting<baylib::bayesian_net<Variable>>(bn, SAMPLES, THREADS).make_inference(),
-                gibbs_sampling<baylib::bayesian_net<Variable>>(bn, SAMPLES, THREADS).make_inference(),
-                rejection_sampling<baylib::bayesian_net<Variable>>(bn, SAMPLES, THREADS).make_inference(),
-                adaptive_importance_sampling<baylib::bayesian_net<Variable>>(bn, SAMPLES, MEMORY).make_inference()
+                likelihood_weighting<bnet<Variable>>(bn, SAMPLES, THREADS).make_inference(),
+                gibbs_sampling<bnet<Variable>>(bn, SAMPLES, THREADS).make_inference(),
+                rejection_sampling<bnet<Variable>>(bn, SAMPLES, THREADS).make_inference(),
+                adaptive_importance_sampling<bnet<Variable>>(bn, SAMPLES, MEMORY).make_inference()
             };
             return results;
         }

diff --git a/test/inference_test.cpp b/test/inference_test.cpp
@@ -22,32 +22,33 @@
 
 using namespace baylib::inference;
 using probability_type = double;
-
+template<class Variable>
+using bnet = baylib::bayesian_net<Variable>;;
 
 template<typename Probability, class Variable>
-std::vector<baylib::marginal_distribution<Probability>> get_results(const baylib::bayesian_net<Variable> &bn){
+std::vector<baylib::marginal_distribution<Probability>> get_results(const bnet<Variable> &bn){
     std::vector<baylib::marginal_distribution<Probability>> results{
-        logic_sampling<baylib::bayesian_net<Variable>>(bn, SAMPLES, MEMORY).make_inference(),
-        gibbs_sampling<baylib::bayesian_net<Variable>>(bn, SAMPLES, THREADS).make_inference(),
-        likelihood_weighting<baylib::bayesian_net<Variable>>(bn, SAMPLES, THREADS).make_inference(),
-        rejection_sampling<baylib::bayesian_net<Variable>>(bn, SAMPLES, THREADS).make_inference(),
-        adaptive_importance_sampling<baylib::bayesian_net<Variable>>(bn, SAMPLES, MEMORY).make_inference(),
+        logic_sampling<bnet<Variable>>(bn, SAMPLES, MEMORY).make_inference(),
+        gibbs_sampling<bnet<Variable>>(bn, SAMPLES, THREADS).make_inference(),
+        likelihood_weighting<bnet<Variable>>(bn, SAMPLES, THREADS).make_inference(),
+        rejection_sampling<bnet<Variable>>(bn, SAMPLES, THREADS).make_inference(),
+        adaptive_importance_sampling<bnet<Variable>>(bn, SAMPLES, MEMORY).make_inference(),
 #ifdef CUDA_CMP_FOUND
-        logic_sampling_cuda<baylib::bayesian_net<Variable>>(bn, SAMPLES).make_inference()
+        logic_sampling_cuda<bnet<Variable>>(bn, SAMPLES).make_inference()
 #endif
     };
     return results;
 }
 
 template<typename Probability, class Variable>
-        std::vector<baylib::marginal_distribution<Probability>> get_results_deterministic(const baylib::bayesian_net<Variable> &bn){
+        std::vector<baylib::marginal_distribution<Probability>> get_results_deterministic(const bnet<Variable> &bn){
             std::vector<baylib::marginal_distribution<Probability>> results{
-                logic_sampling<baylib::bayesian_net<Variable>>(bn, SAMPLES, MEMORY).make_inference(),
-                likelihood_weighting<baylib::bayesian_net<Variable>>(bn, SAMPLES, THREADS).make_inference(),
-                rejection_sampling<baylib::bayesian_net<Variable>>(bn, SAMPLES, THREADS).make_inference(),
-                adaptive_importance_sampling<baylib::bayesian_net<Variable>>(bn, SAMPLES, MEMORY).make_inference(),
+                logic_sampling<bnet<Variable>>(bn, SAMPLES, MEMORY).make_inference(),
+                likelihood_weighting<bnet<Variable>>(bn, SAMPLES, THREADS).make_inference(),
+                rejection_sampling<bnet<Variable>>(bn, SAMPLES, THREADS).make_inference(),
+                adaptive_importance_sampling<bnet<Variable>>(bn, SAMPLES, MEMORY).make_inference(),
 #ifdef CUDA_CMP_FOUND
-                logic_sampling_cuda<baylib::bayesian_net<Variable>>(bn, SAMPLES).make_inference()
+                logic_sampling_cuda<bnet<Variable>>(bn, SAMPLES).make_inference()
 #endif
             };
             return results;