Changing define names.

GEOS-DEV · Aug 12, 2020 · 2bf471c · 2bf471c
1 parent ae95dc9
commit 2bf471c
Show file tree

Hide file tree

Showing 72 changed files with 249 additions and 286 deletions.
diff --git a/benchmarks/benchmarkArray1DR2TensorMultiplication.cpp b/benchmarks/benchmarkArray1DR2TensorMultiplication.cpp
@@ -218,12 +218,12 @@ void pointerRAJA( benchmark::State & state )
 
 INDEX_TYPE const SERIAL_SIZE = (2 << 18) - 87;
 
-#if defined(USE_OPENMP)
+#if defined(LVARRAY_USE_OPENMP)
 INDEX_TYPE const OMP_SIZE = (2 << 22) - 87;
 #endif
 
 // The non Array benchmarks could be run without chai, but then what's the point.
-#if defined(USE_CUDA) && defined(USE_CHAI)
+#if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
 constexpr INDEX_TYPE CUDA_SIZE = (2 << 24) - 87;
 #endif
 
@@ -271,11 +271,11 @@ void registerBenchmarks()
  },
  std::make_tuple( SERIAL_SIZE, RAJA::PERM_IJK {}, serialPolicy {} )
  , std::make_tuple( SERIAL_SIZE, RAJA::PERM_KJI {}, serialPolicy {} )
- #if defined(USE_OPENMP)
+ #if defined(LVARRAY_USE_OPENMP)
  , std::make_tuple( OMP_SIZE, RAJA::PERM_IJK {}, parallelHostPolicy {} )
  , std::make_tuple( OMP_SIZE, RAJA::PERM_KJI {}, parallelHostPolicy {} )
  #endif
- #if defined(USE_CUDA) && defined(USE_CHAI)
+ #if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
  , std::make_tuple( CUDA_SIZE, RAJA::PERM_IJK {}, parallelDevicePolicy< THREADS_PER_BLOCK > {} )
  , std::make_tuple( CUDA_SIZE, RAJA::PERM_KJI {}, parallelDevicePolicy< THREADS_PER_BLOCK > {} )
  #endif
@@ -295,11 +295,11 @@ int main( int argc, char * * argv )
  LVARRAY_LOG( "VALUE_TYPE = " << LvArray::system::demangleType< LvArray::benchmarking::VALUE_TYPE >() );
  LVARRAY_LOG( "Serial problems of size ( " << LvArray::benchmarking::SERIAL_SIZE << ", 3, 3 )." );
 
-#if defined(USE_OPENMP)
+#if defined(LVARRAY_USE_OPENMP)
  LVARRAY_LOG( "OMP problems of size ( " << LvArray::benchmarking::OMP_SIZE << ", 3, 3 )." );
 #endif
 
-#if defined(USE_CUDA) && defined(USE_CHAI)
+#if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
  LVARRAY_LOG( "CUDA problems of size ( " << LvArray::benchmarking::CUDA_SIZE << ", 3, 3 )." );
 #endif
 

diff --git a/benchmarks/benchmarkArray1DR2TensorMultiplicationKernels.cpp b/benchmarks/benchmarkArray1DR2TensorMultiplicationKernels.cpp
@@ -290,12 +290,12 @@ template class ArrayOfR2TensorsNative< RAJA::PERM_KJI >;
 template class ArrayOfR2TensorsRAJA< RAJA::PERM_IJK, serialPolicy >;
 template class ArrayOfR2TensorsRAJA< RAJA::PERM_KJI, serialPolicy >;
 
-#if defined(USE_OPENMP)
+#if defined(LVARRAY_USE_OPENMP)
 template class ArrayOfR2TensorsRAJA< RAJA::PERM_IJK, parallelHostPolicy >;
 template class ArrayOfR2TensorsRAJA< RAJA::PERM_KJI, parallelHostPolicy >;
 #endif
 
-#if defined(USE_CUDA) && defined(USE_CHAI)
+#if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
 template class ArrayOfR2TensorsRAJA< RAJA::PERM_IJK, RAJA::cuda_exec< THREADS_PER_BLOCK > >;
 template class ArrayOfR2TensorsRAJA< RAJA::PERM_KJI, RAJA::cuda_exec< THREADS_PER_BLOCK > >;
 #endif

diff --git a/benchmarks/benchmarkArrayOfArraysNodeToElementMapConstruction.cpp b/benchmarks/benchmarkArrayOfArraysNodeToElementMapConstruction.cpp
@@ -67,7 +67,7 @@ void registerBenchmarks()
  REGISTER_BENCHMARK_TEMPLATE( WRAP( { nx, ny, nz } ), overAllocation, POLICY );
  REGISTER_BENCHMARK_TEMPLATE( WRAP( { nx, ny, nz } ), resizeFromCapacities, POLICY );
  }, std::make_tuple( NX, NY, NZ, serialPolicy {} )
- #if defined(USE_OPENMP)
+ #if defined(LVARRAY_USE_OPENMP)
  , std::make_tuple( NX, NY, NZ, parallelHostPolicy {} )
  #endif
  );

diff --git a/benchmarks/benchmarkArrayOfArraysNodeToElementMapConstructionKernels.cpp b/benchmarks/benchmarkArrayOfArraysNodeToElementMapConstructionKernels.cpp
@@ -127,7 +127,7 @@ resizeFromCapacities( ArrayView< INDEX_TYPE const, 2, 1, INDEX_TYPE, DEFAULT_BUF
 // Explicit instantiation of NodeToElemMapConstruction.
 template class NodeToElemMapConstruction< serialPolicy >;
 
-#if defined(USE_OPENMP)
+#if defined(LVARRAY_USE_OPENMP)
 template class NodeToElemMapConstruction< parallelHostPolicy >;
 #endif
 

diff --git a/benchmarks/benchmarkArrayOfArraysNodeToElementMapConstructionKernels.hpp b/benchmarks/benchmarkArrayOfArraysNodeToElementMapConstructionKernels.hpp
@@ -49,7 +49,7 @@ class NaiveNodeToElemMapConstruction
  {
  CALI_CXX_MARK_SCOPE( "~NaiveNodeToElemMapConstruction" );
 
- // #if defined(USE_OPENMP)
+ // #if defined(LVARRAY_USE_OPENMP)
  // using EXEC_POLICY = parallelHostPolicy;
  // #else
  using EXEC_POLICY = serialPolicy;

diff --git a/benchmarks/benchmarkEigendecomposition.cpp b/benchmarks/benchmarkEigendecomposition.cpp
@@ -38,13 +38,13 @@ void eigenvectors( benchmark::State & state )
 INDEX_TYPE const SERIAL_SIZE_2x2 = (2 << 22) - 87;
 INDEX_TYPE const SERIAL_SIZE_3x3 = (2 << 19) - 87;
 
-#if defined(USE_OPENMP)
+#if defined(LVARRAY_USE_OPENMP)
 INDEX_TYPE const OMP_SIZE_2x2 = (2 << 24) - 87;
 INDEX_TYPE const OMP_SIZE_3x3 = (2 << 23) - 87;
 #endif
 
 // The non Array benchmarks could be run without chai, but then what's the point.
-#if defined(USE_CUDA) && defined(USE_CHAI)
+#if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
 constexpr INDEX_TYPE CUDA_SIZE_2x2 = (2 << 24) - 87;
 constexpr INDEX_TYPE CUDA_SIZE_3x3 = (2 << 24) - 87;
 #endif
@@ -67,13 +67,13 @@ void registerBenchmarks()
  , std::make_tuple( SERIAL_SIZE_2x2, std::integral_constant< int, 2 > {}, RAJA::PERM_JI {}, RAJA::PERM_KJI {}, serialPolicy {} )
  , std::make_tuple( SERIAL_SIZE_3x3, std::integral_constant< int, 3 > {}, RAJA::PERM_IJ {}, RAJA::PERM_IJK {}, serialPolicy {} )
  , std::make_tuple( SERIAL_SIZE_3x3, std::integral_constant< int, 3 > {}, RAJA::PERM_JI {}, RAJA::PERM_KJI {}, serialPolicy {} )
- #if defined(USE_OPENMP)
+ #if defined(LVARRAY_USE_OPENMP)
  , std::make_tuple( OMP_SIZE_2x2, std::integral_constant< int, 2 > {}, RAJA::PERM_IJ {}, RAJA::PERM_IJK {}, parallelHostPolicy {} )
  , std::make_tuple( OMP_SIZE_2x2, std::integral_constant< int, 2 > {}, RAJA::PERM_JI {}, RAJA::PERM_KJI {}, parallelHostPolicy {} )
  , std::make_tuple( OMP_SIZE_3x3, std::integral_constant< int, 3 > {}, RAJA::PERM_IJ {}, RAJA::PERM_IJK {}, parallelHostPolicy {} )
  , std::make_tuple( OMP_SIZE_3x3, std::integral_constant< int, 3 > {}, RAJA::PERM_JI {}, RAJA::PERM_KJI {}, parallelHostPolicy {} )
  #endif
- #if defined(USE_CUDA) && defined(USE_CHAI)
+ #if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
  , std::make_tuple( CUDA_SIZE_2x2, std::integral_constant< int, 2 > {}, RAJA::PERM_IJ {}, RAJA::PERM_IJK {}, parallelDevicePolicy< THREADS_PER_BLOCK > {} )
  , std::make_tuple( CUDA_SIZE_2x2, std::integral_constant< int, 2 > {}, RAJA::PERM_JI {}, RAJA::PERM_KJI {}, parallelDevicePolicy< THREADS_PER_BLOCK > {} )
  , std::make_tuple( CUDA_SIZE_3x3, std::integral_constant< int, 3 > {}, RAJA::PERM_IJ {}, RAJA::PERM_IJK {}, parallelDevicePolicy< THREADS_PER_BLOCK > {} )
@@ -97,12 +97,12 @@ int main( int argc, char * * argv )
  LVARRAY_LOG( "Serial number of 2x2 matrices = " << LvArray::benchmarking::SERIAL_SIZE_2x2 );
  LVARRAY_LOG( "Serial number of 3x3 matrices = " << LvArray::benchmarking::SERIAL_SIZE_3x3 );
 
-#if defined(USE_OPENMP)
+#if defined(LVARRAY_USE_OPENMP)
  LVARRAY_LOG( "OMP number of 2x2 matrices = " << LvArray::benchmarking::OMP_SIZE_2x2 );
  LVARRAY_LOG( "OMP number of 3x3 matrices = " << LvArray::benchmarking::OMP_SIZE_3x3 );
 #endif
 
-#if defined(USE_CUDA) && defined(USE_CHAI)
+#if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
  LVARRAY_LOG( "CUDA number of 2x2 matrices = " << LvArray::benchmarking::CUDA_SIZE_2x2 );
  LVARRAY_LOG( "CUDA number of 3x3 matrices = " << LvArray::benchmarking::CUDA_SIZE_3x3 );
 #endif

diff --git a/benchmarks/benchmarkEigendecompositionKernels.cpp b/benchmarks/benchmarkEigendecompositionKernels.cpp
@@ -41,14 +41,14 @@ template class Eigendecomposition< 2, RAJA::PERM_JI, RAJA::PERM_KJI, serialPolic
 template class Eigendecomposition< 3, RAJA::PERM_IJ, RAJA::PERM_IJK, serialPolicy >;
 template class Eigendecomposition< 3, RAJA::PERM_JI, RAJA::PERM_KJI, serialPolicy >;
 
-#if defined(USE_OPENMP)
+#if defined(LVARRAY_USE_OPENMP)
 template class Eigendecomposition< 2, RAJA::PERM_IJ, RAJA::PERM_IJK, parallelHostPolicy >;
 template class Eigendecomposition< 2, RAJA::PERM_JI, RAJA::PERM_KJI, parallelHostPolicy >;
 template class Eigendecomposition< 3, RAJA::PERM_IJ, RAJA::PERM_IJK, parallelHostPolicy >;
 template class Eigendecomposition< 3, RAJA::PERM_JI, RAJA::PERM_KJI, parallelHostPolicy >;
 #endif
 
-#if defined(USE_CUDA) && defined(USE_CHAI)
+#if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
 template class Eigendecomposition< 2, RAJA::PERM_IJ, RAJA::PERM_IJK, parallelDevicePolicy< THREADS_PER_BLOCK > >;
 template class Eigendecomposition< 2, RAJA::PERM_JI, RAJA::PERM_KJI, parallelDevicePolicy< THREADS_PER_BLOCK > >;
 template class Eigendecomposition< 3, RAJA::PERM_IJ, RAJA::PERM_IJK, parallelDevicePolicy< THREADS_PER_BLOCK > >;

diff --git a/benchmarks/benchmarkHelpers.hpp b/benchmarks/benchmarkHelpers.hpp
@@ -18,7 +18,7 @@
 #include <chrono>
 #include <unordered_map>
 
-#if defined(USE_CALIPER)
+#if defined(LVARRAY_USE_CALIPER)
  #include <caliper/cali.h>
 #define CALI_CXX_MARK_PRETTY_FUNCTION cali::Function __cali_ann ## __func__( __PRETTY_FUNCTION__ )
 #else
@@ -32,7 +32,7 @@ namespace LvArray
 using namespace testing;
 
 
-#if defined(USE_CHAI)
+#if defined(LVARRAY_USE_CHAI)
 static_assert( std::is_same< DEFAULT_BUFFER< int >, ChaiBuffer< int > >::value,
  "The default buffer should be ChaiBuffer when chai is enabled." );
 #endif

diff --git a/benchmarks/benchmarkInnerProduct.cpp b/benchmarks/benchmarkInnerProduct.cpp
@@ -119,12 +119,12 @@ void pointerRAJA( benchmark::State & state )
 
 
 INDEX_TYPE const SERIAL_SIZE = (2 << 20) + 573;
-#if defined(USE_OPENMP)
+#if defined(LVARRAY_USE_OPENMP)
 INDEX_TYPE const OMP_SIZE = SERIAL_SIZE;
 #endif
 
 // The non Array benchmarks could be run without chai, but then what's the point.
-#if defined(USE_CUDA) && defined(USE_CHAI)
+#if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
 INDEX_TYPE const CUDA_SIZE = SERIAL_SIZE;
 #endif
 
@@ -153,10 +153,10 @@ void registerBenchmarks()
  REGISTER_BENCHMARK_TEMPLATE( { size }, pointerRAJA, POLICY );
  },
  std::make_tuple( SERIAL_SIZE, serialPolicy {} )
- #if defined(USE_OPENMP)
+ #if defined(LVARRAY_USE_OPENMP)
  , std::make_tuple( OMP_SIZE, parallelHostPolicy {} )
  #endif
- #if defined(USE_CUDA) && defined(USE_CHAI)
+ #if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
  , std::make_tuple( CUDA_SIZE, parallelDevicePolicy< THREADS_PER_BLOCK > {} )
  #endif
  );
@@ -178,11 +178,11 @@ int main( int argc, char * * argv )
 
  LVARRAY_LOG( "Serial problems of size ( " << LvArray::benchmarking::SERIAL_SIZE << " )." );
 
-#if defined(USE_OPENMP)
+#if defined(LVARRAY_USE_OPENMP)
  LVARRAY_LOG( "OMP problems of size ( " << LvArray::benchmarking::OMP_SIZE << " )." );
 #endif
 
-#if defined(USE_CUDA) && defined(USE_CHAI)
+#if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
  LVARRAY_LOG( "CUDA problems of size ( " << LvArray::benchmarking::CUDA_SIZE << " )." );
 #endif
 

diff --git a/benchmarks/benchmarkInnerProductKernels.cpp b/benchmarks/benchmarkInnerProductKernels.cpp
@@ -109,11 +109,11 @@ pointerKernel( INDEX_TYPE const N,
 
 template class InnerProductRAJA< serialPolicy >;
 
-#if defined(USE_OPENMP)
+#if defined(LVARRAY_USE_OPENMP)
 template class InnerProductRAJA< parallelHostPolicy >;
 #endif
 
-#if defined(USE_CUDA) && defined(USE_CHAI)
+#if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
 template class InnerProductRAJA< RAJA::cuda_exec< THREADS_PER_BLOCK > >;
 #endif
 

diff --git a/benchmarks/benchmarkMatrixMatrix.cpp b/benchmarks/benchmarkMatrixMatrix.cpp
@@ -121,13 +121,13 @@ INDEX_TYPE const SERIAL_N = (2 << 7) + 73;
 INDEX_TYPE const SERIAL_L = (2 << 7) - 71;
 INDEX_TYPE const SERIAL_M = (2 << 7) - 3;
 
-#if defined(USE_OPENMP)
+#if defined(LVARRAY_USE_OPENMP)
 INDEX_TYPE const OMP_N = SERIAL_N;
 INDEX_TYPE const OMP_L = SERIAL_L;
 INDEX_TYPE const OMP_M = SERIAL_M;
 #endif
 
-#if defined(USE_CUDA) && defined(USE_CHAI)
+#if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
 INDEX_TYPE const CUDA_N = SERIAL_N;
 INDEX_TYPE const CUDA_L = SERIAL_L;
 INDEX_TYPE const CUDA_M = SERIAL_M;
@@ -169,11 +169,11 @@ void registerBenchmarks()
  },
  std::make_tuple( SERIAL_N, SERIAL_L, SERIAL_M, RAJA::PERM_IJ {}, serialPolicy {} )
  , std::make_tuple( SERIAL_N, SERIAL_L, SERIAL_M, RAJA::PERM_JI {}, serialPolicy {} )
- #if defined(USE_OPENMP)
+ #if defined(LVARRAY_USE_OPENMP)
  , std::make_tuple( OMP_N, SERIAL_L, OMP_M, RAJA::PERM_IJ {}, parallelHostPolicy {} )
  , std::make_tuple( OMP_N, SERIAL_L, OMP_M, RAJA::PERM_JI {}, parallelHostPolicy {} )
  #endif
- #if defined(USE_CUDA) && defined(USE_CHAI)
+ #if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
  , std::make_tuple( CUDA_N, SERIAL_L, CUDA_M, RAJA::PERM_IJ {}, parallelDevicePolicy< THREADS_PER_BLOCK > {} )
  , std::make_tuple( CUDA_N, SERIAL_L, CUDA_M, RAJA::PERM_JI {}, parallelDevicePolicy< THREADS_PER_BLOCK > {} )
  #endif
@@ -196,12 +196,12 @@ int main( int argc, char * * argv )
  LVARRAY_LOG( "Serial problems of size ( " << LvArray::benchmarking::SERIAL_N << ", " <<
  LvArray::benchmarking::SERIAL_L << ", " << LvArray::benchmarking::SERIAL_M << " )." );
 
-#if defined(USE_OPENMP)
+#if defined(LVARRAY_USE_OPENMP)
  LVARRAY_LOG( "OMP problems of size ( " << LvArray::benchmarking::OMP_N << ", " <<
  LvArray::benchmarking::OMP_L << ", " << LvArray::benchmarking::OMP_M << " )." );
 #endif
 
-#if defined(USE_CUDA) && defined(USE_CHAI)
+#if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
  LVARRAY_LOG( "CUDA problems of size ( " << LvArray::benchmarking::CUDA_N << ", " <<
  LvArray::benchmarking::CUDA_L << ", " << LvArray::benchmarking::CUDA_M << " )." );
 #endif

diff --git a/benchmarks/benchmarkMatrixMatrixKernels.cpp b/benchmarks/benchmarkMatrixMatrixKernels.cpp
@@ -204,14 +204,14 @@ template class MatrixMatrixNative< RAJA::PERM_JI >;
 template class MatrixMatrixRAJA< RAJA::PERM_IJ, serialPolicy >;
 template class MatrixMatrixRAJA< RAJA::PERM_JI, serialPolicy >;
 
-#if defined(USE_OPENMP)
+#if defined(LVARRAY_USE_OPENMP)
 
 template class MatrixMatrixRAJA< RAJA::PERM_IJ, parallelHostPolicy >;
 template class MatrixMatrixRAJA< RAJA::PERM_JI, parallelHostPolicy >;
 
 #endif
 
-#if defined(USE_CUDA) && defined(USE_CHAI)
+#if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
 
 template class MatrixMatrixRAJA< RAJA::PERM_IJ, RAJA::cuda_exec< THREADS_PER_BLOCK > >;
 template class MatrixMatrixRAJA< RAJA::PERM_JI, RAJA::cuda_exec< THREADS_PER_BLOCK > >;

diff --git a/benchmarks/benchmarkMatrixVector.cpp b/benchmarks/benchmarkMatrixVector.cpp
@@ -161,11 +161,11 @@ void registerBenchmarks()
  },
  std::make_tuple( SERIAL_N, SERIAL_M, RAJA::PERM_IJ {}, serialPolicy {} )
  , std::make_tuple( SERIAL_N, SERIAL_M, RAJA::PERM_JI {}, serialPolicy {} )
- #if defined(USE_OPENMP)
+ #if defined(LVARRAY_USE_OPENMP)
  , std::make_tuple( OMP_N, OMP_M, RAJA::PERM_IJ {}, parallelHostPolicy {} )
  , std::make_tuple( OMP_N, OMP_M, RAJA::PERM_JI {}, parallelHostPolicy {} )
  #endif
- #if defined(USE_CUDA) && defined(USE_CHAI)
+ #if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
  , std::make_tuple( CUDA_N, CUDA_M, RAJA::PERM_IJ {}, parallelDevicePolicy< THREADS_PER_BLOCK > {} )
  , std::make_tuple( CUDA_N, CUDA_M, RAJA::PERM_JI {}, parallelDevicePolicy< THREADS_PER_BLOCK > {} )
  #endif
@@ -188,12 +188,12 @@ int main( int argc, char * * argv )
  LVARRAY_LOG( "Serial problems of size ( " << LvArray::benchmarking::SERIAL_N << ", " <<
  LvArray::benchmarking::SERIAL_M << " )." );
 
-#if defined(USE_OPENMP)
+#if defined(LVARRAY_USE_OPENMP)
  LVARRAY_LOG( "OMP problems of size ( " << LvArray::benchmarking::OMP_N << ", " <<
  LvArray::benchmarking::OMP_M << " )." );
 #endif
 
-#if defined(USE_CUDA) && defined(USE_CHAI)
+#if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
  LVARRAY_LOG( "CUDA problems of size ( " << LvArray::benchmarking::CUDA_N << ", " <<
  LvArray::benchmarking::CUDA_M << " )." );
 #endif

diff --git a/benchmarks/benchmarkMatrixVectorKernels.cpp b/benchmarks/benchmarkMatrixVectorKernels.cpp
@@ -171,14 +171,14 @@ template class MatrixVectorNative< RAJA::PERM_JI >;
 template class MatrixVectorRAJA< RAJA::PERM_IJ, serialPolicy >;
 template class MatrixVectorRAJA< RAJA::PERM_JI, serialPolicy >;
 
-#if defined(USE_OPENMP)
+#if defined(LVARRAY_USE_OPENMP)
 
 template class MatrixVectorRAJA< RAJA::PERM_IJ, parallelHostPolicy >;
 template class MatrixVectorRAJA< RAJA::PERM_JI, parallelHostPolicy >;
 
 #endif
 
-#if defined(USE_CUDA) && defined(USE_CHAI)
+#if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
 
 template class MatrixVectorRAJA< RAJA::PERM_IJ, RAJA::cuda_exec< THREADS_PER_BLOCK > >;
 template class MatrixVectorRAJA< RAJA::PERM_JI, RAJA::cuda_exec< THREADS_PER_BLOCK > >;

diff --git a/benchmarks/benchmarkOuterProduct.cpp b/benchmarks/benchmarkOuterProduct.cpp
@@ -120,12 +120,12 @@ void pointerRAJA( benchmark::State & state )
 INDEX_TYPE const SERIAL_N = (2 << 9) + 73;
 INDEX_TYPE const SERIAL_M = (2 << 9) - 71;
 
-#if defined(USE_OPENMP)
+#if defined(LVARRAY_USE_OPENMP)
 INDEX_TYPE const OMP_N = (2 << 9) + 73;
 INDEX_TYPE const OMP_M = (2 << 9) - 71;
 #endif
 
-#if defined(USE_CUDA) && defined(USE_CHAI)
+#if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
 INDEX_TYPE const CUDA_N = (2 << 9) + 73;
 INDEX_TYPE const CUDA_M = (2 << 9) - 71;
 #endif
@@ -165,11 +165,11 @@ void registerBenchmarks()
  },
  std::make_tuple( SERIAL_N, SERIAL_M, RAJA::PERM_IJ {}, serialPolicy {} )
  , std::make_tuple( SERIAL_N, SERIAL_M, RAJA::PERM_JI {}, serialPolicy {} )
- #if defined(USE_OPENMP)
+ #if defined(LVARRAY_USE_OPENMP)
  , std::make_tuple( OMP_N, OMP_M, RAJA::PERM_IJ {}, parallelHostPolicy {} )
  , std::make_tuple( OMP_N, OMP_M, RAJA::PERM_JI {}, parallelHostPolicy {} )
  #endif
- #if defined(USE_CUDA) && defined(USE_CHAI)
+ #if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
  , std::make_tuple( CUDA_N, CUDA_M, RAJA::PERM_IJ {}, parallelDevicePolicy< THREADS_PER_BLOCK > {} )
  , std::make_tuple( CUDA_N, CUDA_M, RAJA::PERM_JI {}, parallelDevicePolicy< THREADS_PER_BLOCK > {} )
  #endif
@@ -192,12 +192,12 @@ int main( int argc, char * * argv )
  LVARRAY_LOG( "Serial problems of size ( " << LvArray::benchmarking::SERIAL_N << ", " <<
  LvArray::benchmarking::SERIAL_M << " )." );
 
-#if defined(USE_OPENMP)
+#if defined(LVARRAY_USE_OPENMP)
  LVARRAY_LOG( "OMP problems of size ( " << LvArray::benchmarking::OMP_N << ", " <<
  LvArray::benchmarking::OMP_M << " )." );
 #endif
 
-#if defined(USE_CUDA) && defined(USE_CHAI)
+#if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
  LVARRAY_LOG( "CUDA problems of size ( " << LvArray::benchmarking::CUDA_N << ", " <<
  LvArray::benchmarking::CUDA_M << " )." );
 #endif

diff --git a/benchmarks/benchmarkOuterProductKernels.cpp b/benchmarks/benchmarkOuterProductKernels.cpp
@@ -170,12 +170,12 @@ template class OuterProductNative< RAJA::PERM_JI >;
 template class OuterProductRAJA< RAJA::PERM_IJ, serialPolicy >;
 template class OuterProductRAJA< RAJA::PERM_JI, serialPolicy >;
 
-#if defined(USE_OPENMP)
+#if defined(LVARRAY_USE_OPENMP)
 template class OuterProductRAJA< RAJA::PERM_IJ, parallelHostPolicy >;
 template class OuterProductRAJA< RAJA::PERM_JI, parallelHostPolicy >;
 #endif
 
-#if defined(USE_CUDA) && defined(USE_CHAI)
+#if defined(LVARRAY_USE_CUDA) && defined(LVARRAY_USE_CHAI)
 template class OuterProductRAJA< RAJA::PERM_IJ, RAJA::cuda_exec< THREADS_PER_BLOCK > >;
 template class OuterProductRAJA< RAJA::PERM_JI, RAJA::cuda_exec< THREADS_PER_BLOCK > >;
 #endif