Update armadillo version 12.8.3

TLCFEM · May 10, 2024 · 536e1a3 · 536e1a3
1 parent e0795d9
commit 536e1a3
Show file tree

Hide file tree

Showing 7 changed files with 47 additions and 10 deletions.
diff --git a/Include/armadillo/armadillo_bits/arma_version.hpp b/Include/armadillo/armadillo_bits/arma_version.hpp
@@ -23,7 +23,7 @@
 
 #define ARMA_VERSION_MAJOR 12
 #define ARMA_VERSION_MINOR 8
-#define ARMA_VERSION_PATCH 2
+#define ARMA_VERSION_PATCH 3
 #define ARMA_VERSION_NAME  "Cortisol Injector"
 
 

diff --git a/Include/armadillo/armadillo_bits/compiler_setup.hpp b/Include/armadillo/armadillo_bits/compiler_setup.hpp
@@ -377,14 +377,11 @@
 
 
 #if defined(__SUNPRO_CC)
-
   // http://www.oracle.com/technetwork/server-storage/solarisstudio/training/index-jsp-141991.html
   // http://www.oracle.com/technetwork/server-storage/solarisstudio/documentation/cplusplus-faq-355066.html
-
   #if (__SUNPRO_CC < 0x5140)
     #error "*** newer compiler required ***"
   #endif
-
 #endif
 
 

diff --git a/Include/armadillo/armadillo_bits/config.hpp b/Include/armadillo/armadillo_bits/config.hpp
@@ -201,6 +201,11 @@
 
 #if defined(ARMA_EXTRA_DEBUG)
   #undef  ARMA_NO_DEBUG
+  #undef  ARMA_DONT_CHECK_NONFINITE
+
+  #undef  ARMA_CHECK_NONFINITE
+  #define ARMA_CHECK_NONFINITE
+
   #undef  ARMA_WARN_LEVEL
   #define ARMA_WARN_LEVEL 3
 #endif

diff --git a/Include/armadillo/armadillo_bits/config.hpp.cmake b/Include/armadillo/armadillo_bits/config.hpp.cmake
@@ -201,6 +201,11 @@
 
 #if defined(ARMA_EXTRA_DEBUG)
   #undef  ARMA_NO_DEBUG
+  #undef  ARMA_DONT_CHECK_NONFINITE
+
+  #undef  ARMA_CHECK_NONFINITE
+  #define ARMA_CHECK_NONFINITE
+
   #undef  ARMA_WARN_LEVEL
   #define ARMA_WARN_LEVEL 3
 #endif

diff --git a/Include/armadillo/armadillo_bits/debug.hpp b/Include/armadillo/armadillo_bits/debug.hpp
@@ -579,7 +579,7 @@ arma_assert_same_size(const uword A_n_rows, const uword A_n_cols, const uword B_
 
 
 
-//! stop if given matrices have different sizes
+//! stop if given matrices do not have the same size
 template<typename eT1, typename eT2>
 arma_hot
 inline
@@ -600,7 +600,7 @@ arma_assert_same_size(const Mat<eT1>& A, const Mat<eT2>& B, const char* x)
 
 
 
-//! stop if given proxies have different sizes
+//! stop if given proxies do not have the same size
 template<typename eT1, typename eT2>
 arma_hot
 inline
@@ -804,7 +804,7 @@ arma_assert_same_size(const uword A_n_rows, const uword A_n_cols, const uword A_
 
 
 
-//! stop if given cubes have different sizes
+//! stop if given cubes do not have the same size
 template<typename eT1, typename eT2>
 arma_hot
 inline
@@ -883,7 +883,7 @@ arma_assert_same_size(const subview_cube<eT>& A, const ProxyCube<T1>& B, const c
 
 
 
-//! stop if given cube proxies have different sizes
+//! stop if given cube proxies do not have the same size
 template<typename eT1, typename eT2>
 arma_hot
 inline

diff --git a/Include/armadillo/armadillo_bits/fft_engine_fftw3.hpp b/Include/armadillo/armadillo_bits/fft_engine_fftw3.hpp
@@ -23,6 +23,13 @@
 
 #if defined(ARMA_USE_FFTW3)
 
+struct fft_engine_fftw3_aux
+  {
+  #if (!defined(ARMA_DONT_USE_STD_MUTEX))
+  static inline std::mutex& get_plan_mutex() { static std::mutex plan_mutex; return plan_mutex; }
+  #endif
+  };
+
 template<typename cx_type, bool inverse>
 class fft_engine_fftw3
   {
@@ -74,7 +81,30 @@ class fft_engine_fftw3
     const int fftw3_flags = fftw3_flag_destroy | fftw3_flag_estimate;
 
     arma_extra_debug_print("fft_engine_fftw3::constructor: generating 1D plan");
-    fftw3_plan = fftw3::plan_dft_1d<cx_type>(N, X_work.memptr(), Y_work.memptr(), fftw3_sign, fftw3_flags);
+
+    // only fftw3::execute() is thread safe, as per FFTW docs:
+    // https://www.fftw.org/fftw3_doc/Thread-safety.html
+
+    #if defined(ARMA_USE_OPENMP)
+      {
+      #pragma omp critical (arma_fft_engine_fftw3)
+        {
+        fftw3_plan = fftw3::plan_dft_1d<cx_type>(N, X_work.memptr(), Y_work.memptr(), fftw3_sign, fftw3_flags);
+        }
+      }
+    #elif (!defined(ARMA_DONT_USE_STD_MUTEX))
+      {
+      std::mutex& plan_mutex = fft_engine_fftw3_aux::get_plan_mutex();
+
+      const std::lock_guard<std::mutex> lock(plan_mutex);
+
+      fftw3_plan = fftw3::plan_dft_1d<cx_type>(N, X_work.memptr(), Y_work.memptr(), fftw3_sign, fftw3_flags);
+      }
+    #else
+      {
+      fftw3_plan = fftw3::plan_dft_1d<cx_type>(N, X_work.memptr(), Y_work.memptr(), fftw3_sign, fftw3_flags);
+      }
+    #endif
 
     if(fftw3_plan == nullptr)  { arma_stop_runtime_error("fft_engine_fftw3::constructor: failed to create plan"); }
     }

diff --git a/README.md b/README.md
@@ -266,7 +266,7 @@ Additional libraries used in **suanPan** are listed as follows.
 - [**VTK**](https://vtk.org/) version 9.2.6
 - [**CUDA**](https://docs.nvidia.com/cuda/cuda-toolkit-release-notes/) version 12.0
 - [**MAGMA**](https://icl.utk.edu/magma/) version 2.8.0
-- [**Armadillo**](http://arma.sourceforge.net/) version 12.8.2
+- [**Armadillo**](http://arma.sourceforge.net/) version 12.8.3
 - [**ensmallen**](https://ensmallen.org/) version 2.20.0
 - [**oneMKL**](https://software.intel.com/content/www/us/en/develop/tools/oneapi/components/onemkl.html) version 2024.1.0
 - [**Catch2**](https://github.com/catchorg/Catch2) version 3.5.4