Commit 2844b5e7 authored by Doak, Peter W.'s avatar Doak, Peter W.
Browse files

remove mixed precision nfft

parent 4cdc3d66

include/dca/math/nfft/kernels_interface.hpp

+3 −3
Original line number Diff line number Diff line
@@ -49,9 +49,9 @@ void initializeNfftHelper(int nb, int nc, const int* add_r, int lda, const int* 
extern template void initializeNfftHelper<double>(int nb, int nc, const int* add_r, int lda, const int* sub_r, int lds,

                          double t0, double delta_t, double t0_window, double delta_t_window,

                          double beta);

extern template void initializeNfftHelper<float>(int nb, int nc, const int* add_r, int lda, const int* sub_r, int lds,

                          float t0, float delta_t, float t0_window, float delta_t_window,

                          float beta);

// extern template void initializeNfftHelper<float>(int nb, int nc, const int* add_r, int lda, const int* sub_r, int lds,

//                           float t0, float delta_t, float t0_window, float delta_t_window,

//                           float beta);

  

}  // namespace details

}  // namespace nfft

src/math/nfft/dnfft_1d_kernels.cu

+13 −13
Original line number Diff line number Diff line
@@ -178,29 +178,29 @@ template void accumulateOnDevice<oversampling, window_sampling, double, double>( 
    const ConfigElem* config_left, const ConfigElem* config_right, const double* tau,

    const double* cubic_coeff, const int size, cudaStream_t stream_);



template void accumulateOnDevice<oversampling, window_sampling, float, float>(

    const float* M, const int ldm, const float sign, float* out, float* out_sqr, const int ldo,

    const ConfigElem* config_left, const ConfigElem* config_right, const float* tau,

    const float* cubic_coeff, const int size, cudaStream_t stream_);

// template void accumulateOnDevice<oversampling, window_sampling, float, float>(

//     const float* M, const int ldm, const float sign, float* out, float* out_sqr, const int ldo,

//     const ConfigElem* config_left, const ConfigElem* config_right, const float* tau,

//     const float* cubic_coeff, const int size, cudaStream_t stream_);

template void accumulateOnDevice<oversampling, window_sampling, std::complex<double>, double>(

    const std::complex<double>* M, const int ldm, const std::complex<double> sign,

    std::complex<double>* out, std::complex<double>* out_sqr, const int ldo,

    const ConfigElem* config_left, const ConfigElem* config_right, const double* tau,

    const double* cubic_coeff, const int size, cudaStream_t stream_);

template void accumulateOnDevice<oversampling, window_sampling, std::complex<float>, float>(

    const std::complex<float>* M, const int ldm, const std::complex<float> sign,

    std::complex<float>* out, std::complex<float>* out_sqr, const int ldo,

    const ConfigElem* config_left, const ConfigElem* config_right, const float* tau,

    const float* cubic_coeff, const int size, cudaStream_t stream_);

// template void accumulateOnDevice<oversampling, window_sampling, std::complex<float>, float>(

//     const std::complex<float>* M, const int ldm, const std::complex<float> sign,

//     std::complex<float>* out, std::complex<float>* out_sqr, const int ldo,

//     const ConfigElem* config_left, const ConfigElem* config_right, const float* tau,

//     const float* cubic_coeff, const int size, cudaStream_t stream_);



template void sum(const double* in, const int ldi, double* out, const int ldo, const int n,

                  const int m, cudaStream_t stream);

template void sum(const float* in, const int ldi, float* out, const int ldo, const int n,

                  const int m, cudaStream_t stream);

// template void sum(const float* in, const int ldi, float* out, const int ldo, const int n,

//                   const int m, cudaStream_t stream);

template void sum(const std::complex<double>* in, const int ldi, std::complex<double>* out,

                  const int ldo, const int n, const int m, cudaStream_t stream);

template void sum(const std::complex<float>* in, const int ldi, std::complex<float>* out,

                  const int ldo, const int n, const int m, cudaStream_t stream);

// template void sum(const std::complex<float>* in, const int ldi, std::complex<float>* out,

//                   const int ldo, const int n, const int m, cudaStream_t stream);



}  // namespace details

}  // namespace nfft

test/unit/phys/dca_step/cluster_solver/shared_tools/accumulation/sp/sp_accumulator_gpu_test.cpp

+1 −1
Original line number Diff line number Diff line
@@ -35,7 +35,7 @@ using McOptions = MockMcOptions<Scalar>; 
//using Scalar = typename dca::config::McOptions::MCScalar;

template <typename SCALAR>

using SpAccumulatorGpuTest = dca::testing::AccumulationTest<SCALAR, 1, 3, 128>;

using TestTypes = ::testing::Types<float, double, std::complex<double>>;

using TestTypes = ::testing::Types<double, std::complex<double>>;

TYPED_TEST_CASE(SpAccumulatorGpuTest, TestTypes);



TYPED_TEST(SpAccumulatorGpuTest, Accumulate) {