hsolver::calc_grad_with_block_op< T, Device > Struct Template Reference

#include <bpcg_kernel_op.h>

Collaboration diagram for hsolver::calc_grad_with_block_op< T, Device >:

Public Types
using	Real = typename GetTypeReal< T >::type
	dot_real_op computes the dot product of the given complex arrays(treated as float arrays). And there's may have MPI communications while enabling planewave parallization strategy.

Public Member Functions
void	operator() (const Real *prec_in, Real *err_out, Real *beta_out, T *psi_out, T *hpsi_out, T *grad_out, T *grad_old_out, const int &n_basis, const int &n_basis_max, const int &n_band)

Member Typedef Documentation

Real

template<typename T , typename Device >

using hsolver::calc_grad_with_block_op< T, Device >::Real = typename GetTypeReal<T>::type

dot_real_op computes the dot product of the given complex arrays(treated as float arrays). And there's may have MPI communications while enabling planewave parallization strategy.

Input Parameters

Parameters

dev	: the type of computing device
A	: input array arr
dim	: size of A
lda	: leading dimention of A
batch	: batch size, the size of the result array res

Returns

res : the result vector T : dot product result

Member Function Documentation

operator()()

template<typename T , typename Device >

void hsolver::calc_grad_with_block_op< T, Device >::operator()	(	const Real *	prec_in,
		Real *	err_out,
		Real *	beta_out,
		T *	psi_out,
		T *	hpsi_out,
		T *	grad_out,
		T *	grad_old_out,
		const int &	n_basis,
		const int &	n_basis_max,
		const int &	n_band
	)

---

The documentation for this struct was generated from the following file:

• /home/runner/work/abacus-develop/abacus-develop/source/source_hsolver/kernels/bpcg_kernel_op.h