math.cu File Reference

#include "math_kernel.h"
#include <device/device_config.h>
#include <device/cuda/check.h>
#include <stdio.h>
#include <stdlib.h>
#include <math/bcknd/device/device_mpi_reduce.h>
#include <math/bcknd/device/device_mpi_op.h>

Go to the source code of this file.

Functions
void	cuda_copy (void *a, void *b, int *n, cudaStream_t strm)
void	cuda_masked_copy (void *a, void *b, void *mask, int *n, int *m, cudaStream_t strm)
void	cuda_masked_gather_copy (void *a, void *b, void *mask, int *n, int *m, cudaStream_t strm)
void	cuda_masked_atomic_reduction (void *a, void *b, void *mask, int *n, int *m, cudaStream_t strm)
void	cuda_masked_scatter_copy (void *a, void *b, void *mask, int *n, int *m, cudaStream_t strm)
void	cuda_cfill_mask (void *a, real *c, int *size, int *mask, int *mask_size, cudaStream_t strm)
void	cuda_rzero (void *a, int *n, cudaStream_t strm)
void	cuda_cmult (void *a, real *c, int *n, cudaStream_t strm)
void	cuda_cmult2 (void *a, void *b, real *c, int *n, cudaStream_t strm)
void	cuda_cdiv (void *a, real *c, int *n, cudaStream_t strm)
void	cuda_cdiv2 (void *a, void *b, real *c, int *n, cudaStream_t strm)
void	cuda_radd (void *a, real *c, int *n, cudaStream_t strm)
void	cuda_cadd2 (void *a, void *b, real *c, int *n, cudaStream_t strm)
void	cuda_cfill (void *a, real *c, int *n, cudaStream_t strm)
void	cuda_add2 (void *a, void *b, int *n, cudaStream_t strm)
void	cuda_add3 (void *a, void *b, void *c, int *n, cudaStream_t strm)
void	cuda_add4 (void *a, void *b, void *c, void *d, int *n, cudaStream_t strm)
void	cuda_add2s1 (void *a, void *b, real *c1, int *n, cudaStream_t strm)
void	cuda_add2s2 (void *a, void *b, real *c1, int *n, cudaStream_t strm)
void	cuda_add2s2_many (void *x, void **p, void *alpha, int *j, int *n, cudaStream_t strm)
void	cuda_addsqr2s2 (void *a, void *b, real *c1, int *n, cudaStream_t strm)
void	cuda_add3s2 (void *a, void *b, void *c, real *c1, real *c2, int *n, cudaStream_t strm)
void	cuda_add4s3 (void *a, void *b, void *c, void *d, real *c1, real *c2, real *c3, int *n, cudaStream_t strm)
void	cuda_add5s4 (void *a, void *b, void *c, void *d, void *e, real *c1, real *c2, real *c3, real *c4, int *n, cudaStream_t strm)
void	cuda_invcol1 (void *a, int *n, cudaStream_t strm)
void	cuda_invcol2 (void *a, void *b, int *n, cudaStream_t strm)
void	cuda_invcol3 (void *a, void *b, void *c, int *n, cudaStream_t strm)
void	cuda_col2 (void *a, void *b, int *n, cudaStream_t strm)
void	cuda_col3 (void *a, void *b, void *c, int *n, cudaStream_t strm)
void	cuda_subcol3 (void *a, void *b, void *c, int *n, cudaStream_t strm)
void	cuda_sub2 (void *a, void *b, int *n, cudaStream_t strm)
void	cuda_sub3 (void *a, void *b, void *c, int *n, cudaStream_t strm)
void	cuda_addcol3 (void *a, void *b, void *c, int *n, cudaStream_t strm)
void	cuda_addcol4 (void *a, void *b, void *c, void *d, int *n, cudaStream_t strm)
void	cuda_addcol3s2 (void *a, void *b, void *c, real *s, int *n, cudaStream_t strm)
void	cuda_vdot3 (void *dot, void *u1, void *u2, void *u3, void *v1, void *v2, void *v3, int *n, cudaStream_t strm)
void	cuda_vcross (void *u1, void *u2, void *u3, void *v1, void *v2, void *v3, void *w1, void *w2, void *w3, int *n, cudaStream_t strm)
void	cuda_redbuf_check_alloc (int nb)
void	cuda_global_reduce_add (real *bufred, void *bufred_d, int n, const cudaStream_t stream)
real	cuda_vlsc3 (void *u, void *v, void *w, int *n, cudaStream_t stream)
real	cuda_glsc3 (void *a, void *b, void *c, int *n, cudaStream_t stream)
void	cuda_glsc3_many (real *h, void *w, void *v, void *mult, int *j, int *n, cudaStream_t stream)
real	cuda_glsc2 (void *a, void *b, int *n, cudaStream_t stream)
real	cuda_glsubnorm2 (void *a, void *b, int *n, cudaStream_t stream)
real	cuda_glsum (void *a, int *n, cudaStream_t stream)
void	cuda_absval (void *a, int *n, cudaStream_t stream)
void	cuda_pwmax_vec2 (void *a, void *b, int *n, cudaStream_t stream)
void	cuda_pwmax_vec3 (void *a, void *b, void *c, int *n, cudaStream_t stream)
void	cuda_pwmax_sca2 (void *a, real *c, int *n, cudaStream_t stream)
void	cuda_pwmax_sca3 (void *a, void *b, real *c, int *n, cudaStream_t stream)
void	cuda_pwmin_vec2 (void *a, void *b, int *n, cudaStream_t stream)
void	cuda_pwmin_vec3 (void *a, void *b, void *c, int *n, cudaStream_t stream)
void	cuda_pwmin_sca2 (void *a, real *c, int *n, cudaStream_t stream)
void	cuda_pwmin_sca3 (void *a, void *b, real *c, int *n, cudaStream_t stream)
void	cuda_iadd (void *a, int *c, int *n, cudaStream_t stream)

Variables
int	red_s = 0
real *	bufred = NULL
void *	bufred_d = NULL

Function Documentation

cuda_absval()

void cuda_absval	(	void *	a,
		int *	n,
		cudaStream_t	stream
	)

Fortran wrapper absval Take the abs value of a vector of length n

Definition at line 840 of file math.cu.

Here is the call graph for this function:

cuda_add2()

void cuda_add2	(	void *	a,
		void *	b,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for add2 Vector addition \( a = a + b \)

Definition at line 245 of file math.cu.

Here is the call graph for this function:

cuda_add2s1()

void cuda_add2s1	(	void *	a,
		void *	b,
		real *	c1,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for add2s1 Vector addition with scalar multiplication \( a = c_1 a + b \) (multiplication on first argument)

Definition at line 290 of file math.cu.

Here is the call graph for this function:

cuda_add2s2()

void cuda_add2s2	(	void *	a,
		void *	b,
		real *	c1,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for add2s2 Vector addition with scalar multiplication \( a = a + c_1 b \) (multiplication on second argument)

Definition at line 306 of file math.cu.

Here is the call graph for this function:

cuda_add2s2_many()

void cuda_add2s2_many	(	void *	x,
		void **	p,
		void *	alpha,
		int *	j,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for add2s2 Vector addition with scalar multiplication \( x = x + c_1 p1 + c_2p2 + ... + c_jpj \) (multiplication on second argument)

Definition at line 323 of file math.cu.

Here is the call graph for this function:

cuda_add3()

void cuda_add3	(	void *	a,
		void *	b,
		void *	c,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for add3 Vector addition \( a = b + c \)

Definition at line 260 of file math.cu.

Here is the call graph for this function:

cuda_add3s2()

void cuda_add3s2	(	void *	a,
		void *	b,
		void *	c,
		real *	c1,
		real *	c2,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for add3s2 Vector addition with scalar multiplication \( a = c_1 b + c_2 c \) (multiplication on second argument)

Definition at line 356 of file math.cu.

Here is the call graph for this function:

cuda_add4()

void cuda_add4	(	void *	a,
		void *	b,
		void *	c,
		void *	d,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for add4 Vector addition \( a = b + c + d \)

Definition at line 274 of file math.cu.

Here is the call graph for this function:

cuda_add4s3()

void cuda_add4s3	(	void *	a,
		void *	b,
		void *	c,
		void *	d,
		real *	c1,
		real *	c2,
		real *	c3,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for add4s3 Vector addition with scalar multiplication \( a = c_1 b + c_2 c + c_3 d\) (multiplication on second argument)

Definition at line 373 of file math.cu.

Here is the call graph for this function:

cuda_add5s4()

void cuda_add5s4	(	void *	a,
		void *	b,
		void *	c,
		void *	d,
		void *	e,
		real *	c1,
		real *	c2,
		real *	c3,
		real *	c4,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for add5s4 Vector addition with scalar multiplication \( a = a + c_1 b + c_2 c + c_3 d + c_4 e\) (multiplication on second argument)

Definition at line 390 of file math.cu.

Here is the call graph for this function:

cuda_addcol3()

void cuda_addcol3	(	void *	a,
		void *	b,
		void *	c,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for addcol3 \( a = a + b * c \)

Definition at line 518 of file math.cu.

Here is the call graph for this function:

cuda_addcol3s2()

void cuda_addcol3s2	(	void *	a,
		void *	b,
		void *	c,
		real *	s,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for addcol3s2 \( a = a + s(b * c) \)

Definition at line 547 of file math.cu.

Here is the call graph for this function:

cuda_addcol4()

void cuda_addcol4	(	void *	a,
		void *	b,
		void *	c,
		void *	d,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for addcol4 \( a = a + b * c * d\)

Definition at line 532 of file math.cu.

Here is the call graph for this function:

cuda_addsqr2s2()

void cuda_addsqr2s2	(	void *	a,
		void *	b,
		real *	c1,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for addsqr2s2 Vector addition with scalar multiplication \( a = a + c_1 (b * b) \) (multiplication on second argument)

Definition at line 340 of file math.cu.

Here is the call graph for this function:

cuda_cadd2()

void cuda_cadd2	(	void *	a,
		void *	b,
		real *	c,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for cadd2 Add a scalar to vector \( a_i = b_i + c \)

Definition at line 215 of file math.cu.

Here is the call graph for this function:

cuda_cdiv()

void cuda_cdiv	(	void *	a,
		real *	c,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for cdiv Division of constant c by array \( a = c / a \)

Definition at line 174 of file math.cu.

Here is the call graph for this function:

cuda_cdiv2()

void cuda_cdiv2	(	void *	a,
		void *	b,
		real *	c,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for cdiv2 Division of constant c by array \( a = c / b \)

Definition at line 187 of file math.cu.

Here is the call graph for this function:

cuda_cfill()

void cuda_cfill	(	void *	a,
		real *	c,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for cfill Set all elements to a constant c \( a = c \)

Definition at line 229 of file math.cu.

Here is the call graph for this function:

cuda_cfill_mask()

void cuda_cfill_mask	(	void *	a,
		real *	c,
		int *	size,
		int *	mask,
		int *	mask_size,
		cudaStream_t	strm
	)

Fortran wrapper for cfill_mask Fill a scalar to vector \( a_i = s, for i \in mask \)

Definition at line 126 of file math.cu.

Here is the call graph for this function:

cuda_cmult()

void cuda_cmult	(	void *	a,
		real *	c,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for cmult Multiplication by constant c \( a = c \cdot a \)

Definition at line 147 of file math.cu.

Here is the call graph for this function:

cuda_cmult2()

void cuda_cmult2	(	void *	a,
		void *	b,
		real *	c,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for cmult2 Multiplication by constant c \( a = c \cdot b \)

Definition at line 160 of file math.cu.

Here is the call graph for this function:

cuda_col2()

void cuda_col2	(	void *	a,
		void *	b,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for col2 Vector multiplication with 2 vectors \( a = a \cdot b \)

Definition at line 448 of file math.cu.

Here is the call graph for this function:

cuda_col3()

void cuda_col3	(	void *	a,
		void *	b,
		void *	c,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for col3 Vector multiplication with 3 vectors \( a = b \cdot c \)

Definition at line 461 of file math.cu.

Here is the call graph for this function:

cuda_copy()

void cuda_copy	(	void *	a,
		void *	b,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for copy Copy a vector \( a = b \)

Definition at line 59 of file math.cu.

Here is the call graph for this function:

cuda_global_reduce_add()

void cuda_global_reduce_add	(	real *	bufred,
		void *	bufred_d,
		int	n,
		const cudaStream_t	stream
	)

Global additive reduction

Definition at line 628 of file math.cu.

Here is the call graph for this function:

Here is the caller graph for this function:

cuda_glsc2()

real cuda_glsc2	(	void *	a,
		void *	b,
		int *	n,
		cudaStream_t	stream
	)

Fortran wrapper glsc2 Weighted inner product \( a^T b c \)

Definition at line 755 of file math.cu.

Here is the call graph for this function:

cuda_glsc3()

real cuda_glsc3	(	void *	a,
		void *	b,
		void *	c,
		int *	n,
		cudaStream_t	stream
	)

Fortran wrapper glsc3 Weighted inner product \( a^T b c \)

Definition at line 696 of file math.cu.

Here is the call graph for this function:

cuda_glsc3_many()

void cuda_glsc3_many	(	real *	h,
		void *	w,
		void *	v,
		void *	mult,
		int *	j,
		int *	n,
		cudaStream_t	stream
	)

Fortran wrapper for doing an reduction to an array Weighted inner product \( w^T v(n,1:j) c \)

Definition at line 723 of file math.cu.

Here is the call graph for this function:

cuda_glsubnorm2()

real cuda_glsubnorm2	(	void *	a,
		void *	b,
		int *	n,
		cudaStream_t	stream
	)

Fortran wrapper glsubnorm Squared Norm of difference \( \| a - b \|_2^2 \)

Definition at line 784 of file math.cu.

Here is the call graph for this function:

cuda_glsum()

real cuda_glsum	(	void *	a,
		int *	n,
		cudaStream_t	stream
	)

Fortran wrapper glsum Sum a vector of length n

Definition at line 813 of file math.cu.

Here is the call graph for this function:

cuda_iadd()

void cuda_iadd	(	void *	a,
		int *	c,
		int *	n,
		cudaStream_t	stream
	)

Fortran wrapper for iadd Add a scalar to vector \( a_i = a_i + c \)

Definition at line 970 of file math.cu.

Here is the call graph for this function:

cuda_invcol1()

void cuda_invcol1	(	void *	a,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for invcol1 Invert a vector \( a = 1 / a \)

Definition at line 407 of file math.cu.

Here is the call graph for this function:

cuda_invcol2()

void cuda_invcol2	(	void *	a,
		void *	b,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for invcol2 Vector division \( a = a / b \)

Definition at line 420 of file math.cu.

Here is the call graph for this function:

cuda_invcol3()

void cuda_invcol3	(	void *	a,
		void *	b,
		void *	c,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for invcol3 Vector division \( a = b / c \)

Definition at line 434 of file math.cu.

Here is the call graph for this function:

cuda_masked_atomic_reduction()

void cuda_masked_atomic_reduction	(	void *	a,
		void *	b,
		void *	mask,
		int *	n,
		int *	m,
		cudaStream_t	strm
	)

Fortran wrapper for masked atomic reduction update a vector \( a += b(mask) \) where mask is not unique

Definition at line 97 of file math.cu.

Here is the call graph for this function:

cuda_masked_copy()

void cuda_masked_copy	(	void *	a,
		void *	b,
		void *	mask,
		int *	n,
		int *	m,
		cudaStream_t	strm
	)

Fortran wrapper for masked copy Copy a vector \( a(mask) = b(mask) \)

Definition at line 67 of file math.cu.

Here is the call graph for this function:

cuda_masked_gather_copy()

void cuda_masked_gather_copy	(	void *	a,
		void *	b,
		void *	mask,
		int *	n,
		int *	m,
		cudaStream_t	strm
	)

Fortran wrapper for masked gather copy Copy a vector \( a(i) = b(mask(i)) \)

Definition at line 82 of file math.cu.

Here is the call graph for this function:

cuda_masked_scatter_copy()

void cuda_masked_scatter_copy	(	void *	a,
		void *	b,
		void *	mask,
		int *	n,
		int *	m,
		cudaStream_t	strm
	)

Fortran wrapper for masked scatter copy Copy a vector \( a(mask(i)) = b(i) \)

Definition at line 111 of file math.cu.

Here is the call graph for this function:

cuda_pwmax_sca2()

void cuda_pwmax_sca2	(	void *	a,
		real *	c,
		int *	n,
		cudaStream_t	stream
	)

Fortran wrapper for pwmax_sca2

Compute the maximum of vector and scalar \( a = \max(a, c) \)

Definition at line 886 of file math.cu.

Here is the call graph for this function:

cuda_pwmax_sca3()

void cuda_pwmax_sca3	(	void *	a,
		void *	b,
		real *	c,
		int *	n,
		cudaStream_t	stream
	)

Fortran wrapper for pwmax_sca3

Compute the maximum of vector and scalar \( a = \max(b, c) \)

Definition at line 900 of file math.cu.

Here is the call graph for this function:

cuda_pwmax_vec2()

void cuda_pwmax_vec2	(	void *	a,
		void *	b,
		int *	n,
		cudaStream_t	stream
	)

Fortran wrapper for pwmax_vec2

Compute the maximum of two vectors \( a = \max(a, b) \)

Definition at line 858 of file math.cu.

Here is the call graph for this function:

cuda_pwmax_vec3()

void cuda_pwmax_vec3	(	void *	a,
		void *	b,
		void *	c,
		int *	n,
		cudaStream_t	stream
	)

Fortran wrapper for pwmax_vec3

Compute the maximum of two vectors \( a = \max(b, c) \)

Definition at line 872 of file math.cu.

Here is the call graph for this function:

cuda_pwmin_sca2()

void cuda_pwmin_sca2	(	void *	a,
		real *	c,
		int *	n,
		cudaStream_t	stream
	)

Fortran wrapper for pwmin_sca2

Compute the minimum of vector and scalar \( a = \min(a, c) \)

Definition at line 942 of file math.cu.

Here is the call graph for this function:

cuda_pwmin_sca3()

void cuda_pwmin_sca3	(	void *	a,
		void *	b,
		real *	c,
		int *	n,
		cudaStream_t	stream
	)

Fortran wrapper for pwmin_sca3

Compute the minimum of vector and scalar \( a = \min(b, c) \)

Definition at line 955 of file math.cu.

Here is the call graph for this function:

cuda_pwmin_vec2()

void cuda_pwmin_vec2	(	void *	a,
		void *	b,
		int *	n,
		cudaStream_t	stream
	)

Fortran wrapper for pwmin_vec2

Compute the minimum of two vectors \( a = \min(a, b) \)

Definition at line 914 of file math.cu.

Here is the call graph for this function:

cuda_pwmin_vec3()

void cuda_pwmin_vec3	(	void *	a,
		void *	b,
		void *	c,
		int *	n,
		cudaStream_t	stream
	)

Fortran wrapper for pwmin_vec3

Compute the minimum of two vectors \( a = \min(b, c) \)

Definition at line 928 of file math.cu.

Here is the call graph for this function:

cuda_radd()

void cuda_radd	(	void *	a,
		real *	c,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for cadd Add a scalar to vector \( a_i = a_i + c \)

Definition at line 201 of file math.cu.

Here is the call graph for this function:

cuda_redbuf_check_alloc()

void cuda_redbuf_check_alloc

(

int

nb

)

Checks and allocates a buffer of size nb*sizeof(real) for reductions

Definition at line 605 of file math.cu.

Here is the call graph for this function:

Here is the caller graph for this function:

cuda_rzero()

void cuda_rzero	(	void *	a,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for rzero Zero a real vector

Definition at line 140 of file math.cu.

Here is the call graph for this function:

Here is the caller graph for this function:

cuda_sub2()

void cuda_sub2	(	void *	a,
		void *	b,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for sub2 Vector subtraction \( a = a - b \)

Definition at line 490 of file math.cu.

Here is the call graph for this function:

cuda_sub3()

void cuda_sub3	(	void *	a,
		void *	b,
		void *	c,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for sub3 Vector subtraction \( a = b - c \)

Definition at line 504 of file math.cu.

Here is the call graph for this function:

cuda_subcol3()

void cuda_subcol3	(	void *	a,
		void *	b,
		void *	c,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for subcol3 Vector multiplication with 3 vectors \( a = a - b \cdot c \)

Definition at line 475 of file math.cu.

Here is the call graph for this function:

cuda_vcross()

void cuda_vcross	(	void *	u1,
		void *	u2,
		void *	u3,
		void *	v1,
		void *	v2,
		void *	v3,
		void *	w1,
		void *	w2,
		void *	w3,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for vcross \( u = v \times w \)

Definition at line 579 of file math.cu.

Here is the call graph for this function:

cuda_vdot3()

void cuda_vdot3	(	void *	dot,
		void *	u1,
		void *	u2,
		void *	u3,
		void *	v1,
		void *	v2,
		void *	v3,
		int *	n,
		cudaStream_t	strm
	)

Fortran wrapper for vdot3 \( dot = u \cdot v \)

Definition at line 562 of file math.cu.

Here is the call graph for this function:

cuda_vlsc3()

real cuda_vlsc3	(	void *	u,
		void *	v,
		void *	w,
		int *	n,
		cudaStream_t	stream
	)

Fortran wrapper vlsc3 Compute multiplication sum \( dot = u \cdot v \cdot w \)

Definition at line 668 of file math.cu.

Here is the call graph for this function:

Variable Documentation

bufred

real* bufred = NULL

Definition at line 599 of file math.cu.

bufred_d

void* bufred_d = NULL

Definition at line 600 of file math.cu.

red_s

int red_s = 0

Definition at line 598 of file math.cu.