Kaldi CUDA矩阵类与CPU矩阵类的互相转换

CUDA类转换为CPU类

示例1

CuMatrix<BaseFloat> A(RandInt(1, 50), RandInt(1, 50));

A.SetRandn();

Matrix<BaseFloat> A2(A);

示例2

CuMatrix<BaseFloat> A(RandInt(1, 50), RandInt(1, 50));

A.SetRandn();

Matrix<BaseFloat> A2();

A.CopyToMat(A2);

代码实现

src/matrix/kaldi-matrix.h

template<typename OtherReal>

explicit Matrix(const CuMatrixBase<OtherReal> &cu,

MatrixTransposeType trans = kNoTrans);

src/cudamatrix/cu-matrix.h

template<typename Real>

template<typename OtherReal>

Matrix<Real>::Matrix(const CuMatrixBase<OtherReal> &M,

MatrixTransposeType trans) {

if (trans == kNoTrans) Init(M.NumRows(), M.NumCols(), kDefaultStride);

else Init(M.NumCols(), M.NumRows(), kDefaultStride);

M.CopyToMat(this, trans);

}

template<typename OtherReal>

void CopyToMat(MatrixBase<OtherReal> *dst,

MatrixTransposeType trans = kNoTrans) const;

  

src/cudamatrix/cu-matrix.cc

template<typename Real>

template<typename OtherReal>

void CuMatrixBase<Real>::CopyToMat(MatrixBase<OtherReal> *dst,

MatrixTransposeType trans) const {

#if HAVE_CUDA == 1

if (CuDevice::Instantiate().Enabled()) {

if (trans == kTrans || sizeof(OtherReal) != sizeof(Real)) {

CuMatrix<OtherReal> this_trans(*this, trans);

this_trans.CopyToMat(dst, kNoTrans);

} else {

KALDI_ASSERT(dst->NumRows() == NumRows() && dst->NumCols() == NumCols());

if (num_rows_ == 0) return;

CuTimer tim;

 

MatrixIndexT src_pitch = stride_*sizeof(Real);

MatrixIndexT dst_pitch = dst->Stride()*sizeof(Real);

MatrixIndexT width = NumCols()*sizeof(Real);

CU_SAFE_CALL(cudaMemcpy2DAsync(dst->Data(), dst_pitch, this->data_,

src_pitch, width, this->num_rows_,

cudaMemcpyDeviceToHost,

cudaStreamPerThread));

CU_SAFE_CALL(cudaStreamSynchronize(cudaStreamPerThread));

CuDevice::Instantiate().AccuProfile("CuMatrix::CopyToMatD2H", tim);

}

} else

#endif

{

dst->CopyFromMat(Mat(), trans);

}

}

  

   

CPU类转换为CUDA类

示例1

Matrix<BaseFloat> A(RandInt(1, 50), RandInt(1, 50));

A.SetRandn();

CuMatrix<BaseFloat> A2(A);

示例2

Matrix<BaseFloat> A(RandInt(1, 50), RandInt(1, 50));

A.SetRandn();

CuMatrix<BaseFloat> A2();

A2.CopyFromMat(A);

代码实现

src/cudamatrix/cu-matrix.h

template<typename OtherReal>

explicit CuMatrix(const MatrixBase<OtherReal> &other,

MatrixTransposeType trans = kNoTrans);

src/cudamatrix/cu-matrix.cc

template<typename Real>

template<typename OtherReal>

CuMatrix<Real>::CuMatrix(const MatrixBase<OtherReal> &other, MatrixTransposeType trans) {

if (trans == kNoTrans)

this->Resize(other.NumRows(), other.NumCols(), kUndefined);

else

this->Resize(other.NumCols(), other.NumRows(), kUndefined);

this->CopyFromMat(other, trans);

}

   

template<typename Real>

void CuMatrixBase<Real>::CopyFromMat(const MatrixBase<Real> &src,

MatrixTransposeType trans) {

#if HAVE_CUDA == 1

if (CuDevice::Instantiate().Enabled()) {

if (trans == kNoTrans) {

KALDI_ASSERT(src.NumRows() == num_rows_ && src.NumCols() == num_cols_);

CuTimer tim;

 

MatrixIndexT dst_pitch = stride_*sizeof(Real);

MatrixIndexT src_pitch = src.Stride()*sizeof(Real);

MatrixIndexT width = src.NumCols()*sizeof(Real);

CU_SAFE_CALL(cudaMemcpy2DAsync(data_, dst_pitch, src.Data(), src_pitch,

width, src.NumRows(), cudaMemcpyHostToDevice,

cudaStreamPerThread));

CU_SAFE_CALL(cudaStreamSynchronize(cudaStreamPerThread));

 

CuDevice::Instantiate().AccuProfile("CuMatrixBase::CopyFromMat(from CPU)", tim);

} else {

CuMatrix<Real> trans_mat(src); // Do the transpose on the GPU board.

this->CopyFromMat(trans_mat, kTrans);

}

} else

#endif

{

Mat().CopyFromMat(src, trans);

}

}