---

CUDAParPrefixOps.cu File Reference

High-level CUDA parallel prefix (scan) operations (sum, min, max, ...) (wrapping CUB, Thrust, etc). More...

#include <stdlib.h>
#include "CUDAParPrefixOps.h"
#include <thrust/scan.h>
#include <thrust/execution_policy.h>

Go to the source code of this file.

Defines
#define	CUERR
#define	INST_DEV_EXCL_SCAN_SUM_TMPSZ(T)   template long dev_excl_scan_sum_tmpsz<T>(T*, long, T*, T);
#define	INST_DEV_EXCL_SCAN_SUM(T)   template void dev_excl_scan_sum<T>(T*, long, T*, void*, long, T);
Functions
template<typename T> long	dev_excl_scan_sum_tmpsz (T *in_d, long nGroups, T *out_d, T ival)
template<typename T> void	dev_excl_scan_sum (T *in_d, long nGroups, T *out_d, void *scanwork_d, long tsz, T ival)
template<typename T> long	dev_incl_scan_sum_tmpsz (T *in_d, long nGroups, T *out_d)
template<typename T> void	dev_incl_scan_sum (T *in_d, long nGroups, T *out_d, void *scanwork_d, long tsz)
template void	dev_incl_scan_sum< float > (float *, long, float *, void *, long)

---

Detailed Description

High-level CUDA parallel prefix (scan) operations (sum, min, max, ...) (wrapping CUB, Thrust, etc).

Definition in file CUDAParPrefixOps.cu.

---

Define Documentation

#define CUERR

Definition at line 40 of file CUDAParPrefixOps.cu.

#define INST_DEV_EXCL_SCAN_SUM (  T    )     template void dev_excl_scan_sum<T>(T*, long, T*, void*, long, T);

Definition at line 181 of file CUDAParPrefixOps.cu.

#define INST_DEV_EXCL_SCAN_SUM_TMPSZ (  T    )     template long dev_excl_scan_sum_tmpsz<T>(T*, long, T*, T);

Definition at line 180 of file CUDAParPrefixOps.cu.

---

Function Documentation

template<typename T> void dev_excl_scan_sum (  T *    in_d, long    nGroups, T *    out_d, void *    scanwork_d, long    tsz, T    ival )

Definition at line 136 of file CUDAParPrefixOps.cu.

template<typename T> long dev_excl_scan_sum_tmpsz (  T *    in_d, long    nGroups, T *    out_d, T    ival )

Definition at line 130 of file CUDAParPrefixOps.cu.

template<typename T> void dev_incl_scan_sum (  T *    in_d, long    nGroups, T *    out_d, void *    scanwork_d, long    tsz )

Definition at line 159 of file CUDAParPrefixOps.cu.

template void dev_incl_scan_sum< float > (  float *   , long   , float *   , void *   , long    )

template<typename T> long dev_incl_scan_sum_tmpsz (  T *    in_d, long    nGroups, T *    out_d )

Definition at line 154 of file CUDAParPrefixOps.cu.

---