Fix Windows build compatibility (MSVC cl.exe and clang-cl)

cmake/gauxc-dep-versions.cmake

@@ -7,14 +7,14 @@ set( GAUXC_CUB_REVISION   1.10.0 )
 set( GAUXC_CUTLASS_REPOSITORY https://github.com/NVIDIA/cutlass.git )
 set( GAUXC_CUTLASS_REVISION v2.10.0 )
 
-set( GAUXC_EXCHCXX_REPOSITORY https://github.com/wavefunction91/ExchCXX.git )
-set( GAUXC_EXCHCXX_REVISION   v1.0.0 )
+set( GAUXC_EXCHCXX_REPOSITORY https://github.com/lorisercole/ExchCXX.git )
+set( GAUXC_EXCHCXX_REVISION   601f72eb668e0721a8452fc3eaff510f431946b0 )
 
 set( GAUXC_GAU2GRID_REPOSITORY https://github.com/dgasmith/gau2grid.git )
 set( GAUXC_GAU2GRID_REVISION   v2.0.6 )
 
-set( GAUXC_INTEGRATORXX_REPOSITORY https://github.com/wavefunction91/IntegratorXX.git )
-set( GAUXC_INTEGRATORXX_REVISION   1369be58d7a3235dac36d75dd964fef058830622 )
+set( GAUXC_INTEGRATORXX_REPOSITORY https://github.com/lorisercole/IntegratorXX.git )
+set( GAUXC_INTEGRATORXX_REVISION   58012a0b32c45f5b403380fab594047dd4587f55 )
 
 set( GAUXC_HIGHFIVE_REPOSITORY https://github.com/highfive-devs/HighFive.git )
 set( GAUXC_HIGHFIVE_REVISION 805f0e13d09b47c4b01d40682621904aa3b31bb8 )

external/gau2grid/generated_source/gau2grid/gau2grid.h

@@ -79,4 +79,4 @@ void gg_collocation_deriv3(int L, const unsigned long npoints, const double* PRA
 #ifdef __cplusplus
 }
 #endif
-#endif /* GAU2GRID_GUARD_H */
+#endif /* GAU2GRID_GUARD_H */

external/gau2grid/generated_source/gau2grid/gau2grid_pragma.h

@@ -96,4 +96,4 @@
     #define PRAGMA_RESTRICT                                  __restrict__
 
 
-#endif
+#endif

external/gau2grid/generated_source/gau2grid_helper.c

@@ -8,6 +8,7 @@
 #include <math.h>
 #if defined(__clang__) && defined(_MSC_VER)
 #include <malloc.h>
+#include <stdlib.h>
 #elif defined __clang__
 #include <mm_malloc.h>
 #elif defined _MSC_VER

include/gauxc/exceptions.hpp

@@ -54,7 +54,7 @@ class generic_gauxc_exception : public std::exception {
         << "  Line     " << line_  << std::endl;
      auto msg = ss.str();
 
-     return strdup( msg.c_str() );
+     return _strdup( msg.c_str() );
   };
 
 public:
@@ -76,8 +76,13 @@ class generic_gauxc_exception : public std::exception {
 
 }
 
+#ifdef _MSC_VER
+#define GAUXC_GENERIC_EXCEPTION( MSG ) \
+  throw generic_gauxc_exception( __FILE__, __FUNCSIG__, __LINE__, MSG )
+#else
 #define GAUXC_GENERIC_EXCEPTION( MSG ) \
   throw generic_gauxc_exception( __FILE__, __PRETTY_FUNCTION__, __LINE__, MSG )
+#endif
 
 #define GAUXC_PIMPL_NOT_INITIALIZED() \
   GAUXC_GENERIC_EXCEPTION("PIMPL NOT INITIALIZED")

include/gauxc/molecule.hpp

@@ -47,7 +47,7 @@ class Molecule : public std::vector<Atom> {
       })->Z;
   }
 
-  bool operator==(const Molecule& other) {
+  bool operator==(const Molecule& other) const {
     if(other.size() != this->size()) return false;
     for( auto i = 0ul; i < this->size(); ++i )
       if( other[i] != operator[](i) ) return false;

include/gauxc/reduction_driver.hpp

@@ -11,6 +11,7 @@
  */
 #pragma once
 #include <memory>
+#include <string>
 #include <gauxc/gauxc_config.hpp>
 #include <gauxc/runtime_environment.hpp>
 #include <typeindex>

src/CMakeLists.txt

@@ -78,6 +78,30 @@ if( GAUXC_CXX_HAS_WSHADOW )
   target_compile_options( gauxc PRIVATE $<$<COMPILE_LANGUAGE:CXX>: -Wshadow> )
 endif()
 
+if(MSVC)
+  if(CMAKE_CXX_COMPILER_ID MATCHES "Clang")
+    target_compile_options( gauxc PRIVATE
+      -Wno-covered-switch-default
+      -Wno-documentation
+      -Wno-documentation-unknown-command
+      -Wno-implicit-int-float-conversion
+      -Wno-language-extension-token
+      -Wno-reserved-identifier
+      -Wno-shorten-64-to-32
+      -Wno-sign-compare
+      -Wno-undef
+    )
+  else()
+    target_compile_options( gauxc PRIVATE
+      /W2
+      /wd4100  # unreferenced parameter
+      /wd4101  # unreferenced local variable
+      /wd4242  # 'identifier': conversion from 'type1' to 'type2', possible loss of data
+      /wd5219  # implicit conversion from 'type-1' to 'type-2', possible loss of data
+    )
+  endif()
+endif()
+
 target_link_libraries( gauxc PUBLIC 
   ExchCXX::ExchCXX 
   IntegratorXX::IntegratorXX 

src/exceptions/cublas_exception.hpp

@@ -96,7 +96,7 @@ class cublas_exception : public std::exception {
 
      auto msg = ss.str();
 
-     return strdup( msg.c_str() );
+     return _strdup( msg.c_str() );
   }
 
 public:

src/exceptions/cuda_exception.hpp

@@ -48,7 +48,7 @@ class cuda_exception : public std::exception {
 
      auto msg = ss.str();
 
-     return strdup( msg.c_str() );
+     return _strdup( msg.c_str() );
   }
 
 public:

src/exceptions/cutlass_exception.hpp

@@ -48,7 +48,7 @@ class cutlass_exception : public std::exception {
 
      auto msg = ss.str();
 
-     return strdup( msg.c_str() );
+     return _strdup( msg.c_str() );
   }
 
 public:

src/exceptions/hip_exception.hpp

@@ -48,7 +48,7 @@ class hip_exception : public std::exception {
 
      auto msg = ss.str();
 
-     return strdup( msg.c_str() );
+     return _strdup( msg.c_str() );
   }
 
 public:

src/exceptions/hipblas_exception.hpp

@@ -103,7 +103,7 @@ class hipblas_exception : public std::exception {
 
      auto msg = ss.str();
 
-     return strdup( msg.c_str() );
+     return _strdup( msg.c_str() );
   }
 
 public:

src/exceptions/magma_exception.hpp

@@ -46,7 +46,7 @@ class magma_exception : public std::exception {
 
      auto msg = ss.str();
 
-     return strdup( msg.c_str() );
+     return _strdup( msg.c_str() );
   }
 
 public:

src/external/CMakeLists.txt

@@ -21,19 +21,42 @@ if( GAUXC_ENABLE_HDF5 )
       message(STATUS "HighFive REV  = ${GAUXC_HIGHFIVE_REVISION}  ")
       FetchContent_Declare( HighFive
         GIT_REPOSITORY ${GAUXC_HIGHFIVE_REPOSITORY}
-        GIT_TAG        ${GAUXC_HIGHFIVE_REVISION}  
+        GIT_TAG        ${GAUXC_HIGHFIVE_REVISION}
       )
-    
+
       set(HIGHFIVE_USE_BOOST OFF CACHE BOOL "" )
       set(HIGHFIVE_UNIT_TESTS OFF CACHE BOOL "" )
       set(HIGHFIVE_EXAMPLES OFF CACHE BOOL "" )
       #set(HIGHFIVE_PARALLEL_HDF5 ON CACHE BOOL "" )
       set(HIGHFIVE_BUILD_DOCS OFF CACHE BOOL "" )
       FetchContent_MakeAvailable( HighFive )
-
+
+      # HighFive propagates HDF5_DEFINITIONS via its libdeps target.
+      # CMake's FindHDF5 module sets H5_BUILT_AS_DYNAMIC_LIB on Windows
+      # when HDF5_USE_STATIC_LIBRARIES is not set, even for static libs.
+      # Correct this when linking statically.
+      if(WIN32 AND HDF5_PROVIDES_STATIC_LIBS AND TARGET libdeps)
+        get_target_property(_libdeps_defs libdeps INTERFACE_COMPILE_DEFINITIONS)
+        if(_libdeps_defs)
+          list(REMOVE_ITEM _libdeps_defs "H5_BUILT_AS_DYNAMIC_LIB")
+          list(APPEND _libdeps_defs "H5_BUILT_AS_STATIC_LIB")
+          set_target_properties(libdeps PROPERTIES INTERFACE_COMPILE_DEFINITIONS "${_libdeps_defs}")
+        endif()
+      endif()
+
     endif()
     target_sources( gauxc PRIVATE hdf5_write.cxx hdf5_read.cxx )
     target_link_libraries( gauxc PUBLIC HighFive )
+
+    # When linking HDF5 statically on Windows, HDF5's transitive
+    # dependencies (zlib, szip/aec, shlwapi) must be linked explicitly.
+    if(WIN32 AND HDF5_PROVIDES_STATIC_LIBS)
+      find_library(ZLIB_LIBRARY NAMES zlib z)
+      find_library(SZIP_LIBRARY NAMES szip-static szip sz)
+      find_library(AEC_LIBRARY  NAMES aec-static aec)
+      target_link_libraries( gauxc PUBLIC
+        ${ZLIB_LIBRARY} ${SZIP_LIBRARY} ${AEC_LIBRARY} shlwapi )
+    endif()
   else()
     message(WARNING "GAUXC_ENABLE_HDF5 was enabled, but HDF5 was not found, Disabling HDF5 Bindings")
   endif()

src/xc_integrator/integrator_util/exx_screening.cxx

@@ -13,6 +13,7 @@
 #include "host/blas.hpp"
 #include <gauxc/util/div_ceil.hpp>
 #include <chrono>
+#include <bit>
 //#include <mpi.h>
 //#include <fstream>
 #ifdef GAUXC_HAS_CUDA
@@ -195,7 +196,7 @@ void exx_ek_screening(
     }
 
     uint32_t total_shells = 0;
-    for( auto x : task_ek_shells ) total_shells += __builtin_popcount(x);
+    for( auto x : task_ek_shells ) total_shells += std::popcount(x);
 
     std::vector<uint32_t> ek_shells; ek_shells.reserve(total_shells);
     for( auto i_block = 0u; i_block < util::div_ceil(nshells,32); ++i_block ) {

src/xc_integrator/integrator_util/spherical_harmonics.cxx

@@ -168,4 +168,4 @@ void scaled_ylm_matrix(const int lmax, const double* points, const int32_t  npts
     const std::array<double, 3> x = {points[3 * i], points[3 * i + 1], points[3 * i + 2]};
     scaled_ylm_new(lmax, x, center, radius, nlm, ylm_matrix + i * nharmonics);
   }
-}
+}

src/xc_integrator/integrator_util/spherical_harmonics.hpp

@@ -4,4 +4,4 @@
 #include <cassert>
 
 
-void scaled_ylm_matrix(const int lmax, const double* points, const int32_t  npts, const std::array<double, 3> center, const double radius, double* ylm_matrix);
+void scaled_ylm_matrix(const int lmax, const double* points, const int32_t  npts, const std::array<double, 3> center, const double radius, double* ylm_matrix);

src/xc_integrator/local_work_driver/host/obara_saika/src/integral_0.cxx

@@ -30,7 +30,7 @@ void integral_0(size_t npts,
                int ldG, 
                double *weights,
                double *boys_table) {
-   __attribute__((__aligned__(64))) double buffer[1 * NPTS_LOCAL + 3 * NPTS_LOCAL];
+   alignas(64) double buffer[1 * NPTS_LOCAL + 3 * NPTS_LOCAL];
 
    double *temp       = (buffer + 0);
    double *Tval       = (buffer + 1 * NPTS_LOCAL + 0 * NPTS_LOCAL);

src/xc_integrator/local_work_driver/host/obara_saika/src/integral_0_0.cxx

@@ -32,7 +32,7 @@ void integral_0_0(size_t npts,
                   int ldG, 
                   double *weights,
                   double * /*boys_table*/) {
-   __attribute__((__aligned__(64))) double buffer[1 * NPTS_LOCAL + 3 * NPTS_LOCAL];
+   alignas(64) double buffer[1 * NPTS_LOCAL + 3 * NPTS_LOCAL];
 
    double *temp       = (buffer + 0);
    double *Tval       = (buffer + 1 * NPTS_LOCAL + 0 * NPTS_LOCAL);

src/xc_integrator/local_work_driver/host/obara_saika/src/integral_1.cxx

@@ -30,7 +30,7 @@ void integral_1(size_t npts,
                int ldG, 
                double *weights,
                double *boys_table) {
-   __attribute__((__aligned__(64))) double buffer[9 * NPTS_LOCAL + 3 * NPTS_LOCAL];
+   alignas(64) double buffer[9 * NPTS_LOCAL + 3 * NPTS_LOCAL];
 
    double *temp       = (buffer + 0);
    double *Tval       = (buffer + 9 * NPTS_LOCAL + 0 * NPTS_LOCAL);

src/xc_integrator/local_work_driver/host/obara_saika/src/integral_1_0.cxx

@@ -32,12 +32,12 @@ void integral_1_0(size_t npts,
                   int ldG, 
                   double *weights,
                   double *boys_table) {
-   __attribute__((__aligned__(64))) double buffer[3 * NPTS_LOCAL + 3 * NPTS_LOCAL];
+   alignas(64) double buffer[3 * NPTS_LOCAL + 3 * NPTS_LOCAL];
 
-   double * __restrict__ temp       = (buffer + 0);
-   double * __restrict__ Tval       = (buffer + 3 * NPTS_LOCAL + 0 * NPTS_LOCAL);
-   double * __restrict__ Tval_inv_e = (buffer + 3 * NPTS_LOCAL + 1 * NPTS_LOCAL);
-   double * __restrict__ FmT        = (buffer + 3 * NPTS_LOCAL + 2 * NPTS_LOCAL);
+   double * __restrict temp       = (buffer + 0);
+   double * __restrict Tval       = (buffer + 3 * NPTS_LOCAL + 0 * NPTS_LOCAL);
+   double * __restrict Tval_inv_e = (buffer + 3 * NPTS_LOCAL + 1 * NPTS_LOCAL);
+   double * __restrict FmT        = (buffer + 3 * NPTS_LOCAL + 2 * NPTS_LOCAL);
 
    size_t npts_upper = NPTS_LOCAL * (npts / NPTS_LOCAL);
    size_t p_outer = 0;

src/xc_integrator/local_work_driver/host/obara_saika/src/integral_1_1.cxx

@@ -32,7 +32,7 @@ void integral_1_1(size_t npts,
                   int ldG, 
                   double *weights,
                   double *boys_table) {
-   __attribute__((__aligned__(64))) double buffer[9 * NPTS_LOCAL + 3 * NPTS_LOCAL];
+   alignas(64) double buffer[9 * NPTS_LOCAL + 3 * NPTS_LOCAL];
 
    double *temp       = (buffer + 0);
    double *Tval       = (buffer + 9 * NPTS_LOCAL + 0 * NPTS_LOCAL);

src/xc_integrator/local_work_driver/host/obara_saika/src/integral_2.cxx

@@ -30,7 +30,7 @@ void integral_2(size_t npts,
                int ldG, 
                double *weights,
                double *boys_table) {
-   __attribute__((__aligned__(64))) double buffer[31 * NPTS_LOCAL + 3 * NPTS_LOCAL];
+   alignas(64) double buffer[31 * NPTS_LOCAL + 3 * NPTS_LOCAL];
 
    double *temp       = (buffer + 0);
    double *Tval       = (buffer + 31 * NPTS_LOCAL + 0 * NPTS_LOCAL);

src/xc_integrator/local_work_driver/host/obara_saika/src/integral_2_0.cxx

@@ -32,7 +32,7 @@ void integral_2_0(size_t npts,
                   int ldG, 
                   double *weights,
                   double *boys_table) {
-   __attribute__((__aligned__(64))) double buffer[6 * NPTS_LOCAL + 3 * NPTS_LOCAL];
+   alignas(64) double buffer[6 * NPTS_LOCAL + 3 * NPTS_LOCAL];
 
    double *temp       = (buffer + 0);
    double *Tval       = (buffer + 6 * NPTS_LOCAL + 0 * NPTS_LOCAL);

src/xc_integrator/local_work_driver/host/obara_saika/src/integral_2_1.cxx

@@ -32,7 +32,7 @@ void integral_2_1(size_t npts,
                   int ldG, 
                   double *weights,
                   double *boys_table) {
-   __attribute__((__aligned__(64))) double buffer[16 * NPTS_LOCAL + 3 * NPTS_LOCAL];
+   alignas(64) double buffer[16 * NPTS_LOCAL + 3 * NPTS_LOCAL];
 
    double *temp       = (buffer + 0);
    double *Tval       = (buffer + 16 * NPTS_LOCAL + 0 * NPTS_LOCAL);

src/xc_integrator/local_work_driver/host/obara_saika/src/integral_2_2.cxx

@@ -32,7 +32,7 @@ void integral_2_2(size_t npts,
                   int ldG, 
                   double *weights,
                   double *boys_table) {
-   __attribute__((__aligned__(64))) double buffer[31 * NPTS_LOCAL + 3 * NPTS_LOCAL];
+   alignas(64) double buffer[31 * NPTS_LOCAL + 3 * NPTS_LOCAL];
 
    double *temp       = (buffer + 0);
    double *Tval       = (buffer + 31 * NPTS_LOCAL + 0 * NPTS_LOCAL);

src/xc_integrator/local_work_driver/host/obara_saika/src/integral_3.cxx

@@ -30,7 +30,7 @@ void integral_3(size_t npts,
                int ldG, 
                double *weights,
                double *boys_table) {
-   __attribute__((__aligned__(64))) double buffer[74 * NPTS_LOCAL + 3 * NPTS_LOCAL];
+   alignas(64) double buffer[74 * NPTS_LOCAL + 3 * NPTS_LOCAL];
 
    double *temp       = (buffer + 0);
    double *Tval       = (buffer + 74 * NPTS_LOCAL + 0 * NPTS_LOCAL);

src/xc_integrator/local_work_driver/host/obara_saika/src/integral_3_0.cxx

@@ -32,7 +32,7 @@ void integral_3_0(size_t npts,
                   int ldG, 
                   double *weights,
                   double *boys_table) {
-   __attribute__((__aligned__(64))) double buffer[10 * NPTS_LOCAL + 3 * NPTS_LOCAL];
+   alignas(64) double buffer[10 * NPTS_LOCAL + 3 * NPTS_LOCAL];
 
    double *temp       = (buffer + 0);
    double *Tval       = (buffer + 10 * NPTS_LOCAL + 0 * NPTS_LOCAL);

src/xc_integrator/local_work_driver/host/obara_saika/src/integral_3_1.cxx

@@ -32,7 +32,7 @@ void integral_3_1(size_t npts,
                   int ldG, 
                   double *weights,
                   double *boys_table) {
-   __attribute__((__aligned__(64))) double buffer[25 * NPTS_LOCAL + 3 * NPTS_LOCAL];
+   alignas(64) double buffer[25 * NPTS_LOCAL + 3 * NPTS_LOCAL];
 
    double *temp       = (buffer + 0);
    double *Tval       = (buffer + 25 * NPTS_LOCAL + 0 * NPTS_LOCAL);

src/xc_integrator/local_work_driver/host/obara_saika/src/integral_3_2.cxx

@@ -32,7 +32,7 @@ void integral_3_2(size_t npts,
                   int ldG, 
                   double *weights,
                   double *boys_table) {
-   __attribute__((__aligned__(64))) double buffer[46 * NPTS_LOCAL + 3 * NPTS_LOCAL];
+   alignas(64) double buffer[46 * NPTS_LOCAL + 3 * NPTS_LOCAL];
 
    double *temp       = (buffer + 0);
    double *Tval       = (buffer + 46 * NPTS_LOCAL + 0 * NPTS_LOCAL);

src/xc_integrator/local_work_driver/host/obara_saika/src/integral_3_3.cxx

@@ -32,7 +32,7 @@ void integral_3_3(size_t npts,
                   int ldG, 
                   double *weights,
                   double *boys_table) {
-   __attribute__((__aligned__(64))) double buffer[74 * NPTS_LOCAL + 3 * NPTS_LOCAL];
+   alignas(64) double buffer[74 * NPTS_LOCAL + 3 * NPTS_LOCAL];
 
    double *temp       = (buffer + 0);
    double *Tval       = (buffer + 74 * NPTS_LOCAL + 0 * NPTS_LOCAL);

src/xc_integrator/local_work_driver/host/obara_saika/src/integral_4.cxx

@@ -30,7 +30,7 @@ void integral_4(size_t npts,
                int ldG, 
                double *weights,
                double *boys_table) {
-   __attribute__((__aligned__(64))) double buffer[145 * NPTS_LOCAL + 3 * NPTS_LOCAL];
+   alignas(64) double buffer[145 * NPTS_LOCAL + 3 * NPTS_LOCAL];
 
    double *temp       = (buffer + 0);
    double *Tval       = (buffer + 145 * NPTS_LOCAL + 0 * NPTS_LOCAL);

src/xc_integrator/local_work_driver/host/obara_saika/src/integral_4_0.cxx

@@ -32,7 +32,7 @@ void integral_4_0(size_t npts,
                   int ldG, 
                   double *weights,
                   double *boys_table) {
-   __attribute__((__aligned__(64))) double buffer[15 * NPTS_LOCAL + 3 * NPTS_LOCAL];
+   alignas(64) double buffer[15 * NPTS_LOCAL + 3 * NPTS_LOCAL];
 
    double *temp       = (buffer + 0);
    double *Tval       = (buffer + 15 * NPTS_LOCAL + 0 * NPTS_LOCAL);