void
xSyr2k<cl_float2>::
roundtrip_func()
{
timer.Start(timer_id);
cl_int err;
buffer_.buf_a_ = clCreateBuffer(ctx_, CL_MEM_READ_ONLY,
(buffer_.lda_ * buffer_.a_num_vectors_ +
buffer_.offA_) * sizeof(cl_float2),
NULL, &err);
buffer_.buf_b_ = clCreateBuffer(ctx_, CL_MEM_READ_ONLY,
(buffer_.ldb_ * buffer_.b_num_vectors_ +
buffer_.offB_) * sizeof(cl_float2),
NULL, &err);
buffer_.buf_c_ = clCreateBuffer(ctx_, CL_MEM_READ_WRITE,
(buffer_.ldc_ * buffer_.c_num_vectors_ +
buffer_.offC_) * sizeof(cl_float2),
NULL, &err);
this->initialize_gpu_buffer();
clblasCsyr2k(order_, buffer_.uplo_, buffer_.trans_, buffer_.n_,
buffer_.k_, buffer_.alpha_, buffer_.buf_a_, buffer_.offA_,
buffer_.lda_, buffer_.buf_b_, buffer_.offB_, buffer_.ldb_,
buffer_.beta_, buffer_.buf_c_, buffer_.offC_,
buffer_.ldc_, 1, &queue_, 0, NULL, NULL);
err = clEnqueueReadBuffer(queue_, buffer_.buf_c_, CL_TRUE,
buffer_.offC_ * sizeof(cl_float2),
buffer_.ldc_ * buffer_.c_num_vectors_ *
sizeof(cl_float2),
buffer_.c_, 0, NULL, &event_);
clWaitForEvents(1, &event_);
timer.Stop(timer_id);
}
void
xSyr2k<cl_float2>::
call_func()
{
timer.Start(timer_id);
clblasCsyr2k(order_, buffer_.uplo_, buffer_.trans_, buffer_.n_,
buffer_.k_, buffer_.alpha_, buffer_.buf_a_, buffer_.offA_,
buffer_.lda_, buffer_.buf_b_, buffer_.offB_, buffer_.ldb_,
buffer_.beta_, buffer_.buf_c_, buffer_.offC_,
buffer_.ldc_, 1, &queue_, 0, NULL, &event_);
clWaitForEvents(1, &event_);
timer.Stop(timer_id);
}
/** Perform symmetric rank-2k update.
\f$ C = \alpha A B^T + \alpha B A^T \beta C \f$ (trans == MagmaNoTrans), or \n
\f$ C = \alpha A^T B + \alpha B^T A \beta C \f$ (trans == MagmaTrans), \n
where \f$ C \f$ is symmetric.
@param[in]
uplo Whether the upper or lower triangle of C is referenced.
@param[in]
trans Operation to perform on A and B.
@param[in]
n Number of rows and columns of C. n >= 0.
@param[in]
k Number of columns of A and B (for MagmaNoTrans) or rows of A and B (for MagmaTrans). k >= 0.
@param[in]
alpha Scalar \f$ \alpha \f$
@param[in]
dA COMPLEX array on GPU device.
If trans == MagmaNoTrans, the n-by-k matrix A of dimension (ldda,k), ldda >= max(1,n); \n
otherwise, the k-by-n matrix A of dimension (ldda,n), ldda >= max(1,k).
@param[in]
ldda Leading dimension of dA.
@param[in]
dB COMPLEX array on GPU device.
If trans == MagmaNoTrans, the n-by-k matrix B of dimension (lddb,k), lddb >= max(1,n); \n
otherwise, the k-by-n matrix B of dimension (lddb,n), lddb >= max(1,k).
@param[in]
lddb Leading dimension of dB.
@param[in]
beta Scalar \f$ \beta \f$
@param[in,out]
dC COMPLEX array on GPU device.
The n-by-n symmetric matrix C of dimension (lddc,n), lddc >= max(1,n).
@param[in]
lddc Leading dimension of dC.
@ingroup magma_cblas3
*/
extern "C" void
magma_csyr2k(
magma_uplo_t uplo, magma_trans_t trans,
magma_int_t n, magma_int_t k,
magmaFloatComplex alpha,
magmaFloatComplex_const_ptr dA, size_t dA_offset, magma_int_t ldda,
magmaFloatComplex_const_ptr dB, size_t dB_offset, magma_int_t lddb,
magmaFloatComplex beta,
magmaFloatComplex_ptr dC, size_t dC_offset, magma_int_t lddc,
magma_queue_t queue )
{
cl_int err = clblasCsyr2k(
clblasColumnMajor,
clblas_uplo_const( uplo ),
clblas_trans_const( trans ),
n, k,
alpha, dA, dA_offset, ldda,
dB, dB_offset, lddb,
beta, dC, dC_offset, lddc,
1, &queue, 0, NULL, g_event );
check_error( err );
}