vl_rand.h File Reference

#include "vl_numtypes.h"
#include "vl_memory.h"

Include dependency graph for vl_rand.h:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Macros
#define	VL_H_RANDOM_SPLIT(randPtr, resultPtr)   (*((vl_ularge_t*)(resultPtr)) = vlRandNext(randPtr))
	This macro distributes random bytes to the specified pointer.
#define	vlRandFx4(randPtr, resultPtr)
	Generates four random floats.

Typedefs
typedef vl_ularge_t	vl_rand
	Random state type.

Functions
vl_rand	vlRandInit ()
	Initializes a random state (aka, seed) based on the current time.
vl_rand	vlRandNext (vl_rand *rand)
	Returns the "next" 64-bit integer.
void	vlRandFill (vl_rand *rand, void *mem, vl_ularge_t len)
	Fills the specified region of memory with random bytes.
vl_float32_t	vlRandF (vl_rand *randPtr)
	Generate a random float.
void	vlRandFx2 (vl_rand *randPtr, vl_float32_t *resultPtr)
	Generates two random floats.
vl_float64_t	vlRandD (vl_rand *randPtr)
	Generate a random double.

Macro Definition Documentation

VL_H_RANDOM_SPLIT

#define VL_H_RANDOM_SPLIT	(		randPtr,
			resultPtr )   (*((vl_ularge_t*)(resultPtr)) = vlRandNext(randPtr))

This macro distributes random bytes to the specified pointer.

This macro results in a reinterpretation suitable for all integer types, but not for floating point types.

The total number of bytes distributed is equal to sizeof(vl_ularge_t).

vlRandFx4

#define vlRandFx4	(		randPtr,
			resultPtr )

Value:

                                        vlRandFx2(randPtr, (vl_float32_t*)(resultPtr));\
                                        vlRandFx2(randPtr, ((vl_float32_t*)(resultPtr)) + 2)

Generates four random floats.

Range of returned floats is in range of 0...1.

Parameters

randPtr	pointer to random state
resultPtr	where the result will be stored.

Complexity of O(1) constant.

Returns

void

Typedef Documentation

vl_rand

typedef vl_ularge_t vl_rand

Random state type.

The PRNG algorithm used by the VL library is splitmix64 or splitmix32. It offers fast pseudo-random number generation and passes Big Crush with minimal state. All vlRand* functions are almost entirely branchless.

Function Documentation

vlRandD()

vl_float64_t vlRandD

(

vl_rand *

randPtr

)

Generate a random double.

Range of returned value is in 0...1.

Parameters

randPtr

pointer to random state

Complexity of O(1) constant.

Returns

random double.

vlRandF()

vl_float32_t vlRandF

(

vl_rand *

randPtr

)

Generate a random float.

Range of returned value is in 0...1.

Parameters

randPtr

pointer to random state

Complexity of O(1) constant.

Returns

random float

Here is the call graph for this function:

Here is the caller graph for this function:

vlRandFill()

void vlRandFill	(	vl_rand *	rand,
		void *	mem,
		vl_ularge_t	len )

Fills the specified region of memory with random bytes.

This function operates on blocks of 8 bytes at a time, until it fills in the last few remaining bytes.

Parameters

rand	rand state
mem	destination
len	total number of bytes to fill.

Complexity of O(n) linear.

vlRandFx2()

void vlRandFx2	(	vl_rand *	randPtr,
		vl_float32_t *	resultPtr )

Generates two random floats.

Range of returned floats is in range of 0...1.

Parameters

randPtr	pointer to random state
resultPtr	where the result will be stored

Complexity of O(1) constant.

Here is the call graph for this function:

vlRandInit()

vl_rand vlRandInit

(

)

Initializes a random state (aka, seed) based on the current time.

This function can be skipped entirely, assuming another seed integer can be provided.

Complexity of O(1) constant.

Returns

random state/seed.

Here is the call graph for this function:

vlRandNext()

vl_rand vlRandNext

(

vl_rand *

rand

)

Returns the "next" 64-bit integer.

This function has a side effect of updating the random state.

Complexity of O(1) constant.

Parameters

rand	pointer to random state

Returns

unsigned 64-bit integer

Here is the caller graph for this function: