bithacks Namespace Reference

Functions
unsigned int	log2_32 (uint32_t v)
unsigned int	log2_64 (uint64_t v)
template<class Int>
unsigned int	log2 (Int v)
uint32_t	ceil_pow2_32 (uint32_t v)
uint64_t	ceil_pow2_64 (uint64_t v)
template<class Int>
unsigned int	ceil_pow2 (Int v)
int	log2_norm (float v)
int	log2_norm (double v)

---

Detailed Description

All the code here is adapted from Sean Anderson's bit twiddling hacks page. http://graphics.stanford.edu/~seander/bithacks.html

---

Function Documentation

template<class Int> unsigned int bithacks::ceil_pow2 (  Int  v  )

uint32_t bithacks::ceil_pow2_32 (  uint32_t  v  )  [inline]

uint64_t bithacks::ceil_pow2_64 (  uint64_t  v  )  [inline]

template<class Int> unsigned int bithacks::log2 (  Int  v  )

unsigned int bithacks::log2_32 (  uint32_t  v  )  [inline]

unsigned int bithacks::log2_64 (  uint64_t  v  )  [inline]

int bithacks::log2_norm (  double  v  )  [inline]

Return the exponent of a normalized double. On a denorm, this returns the bias of -1023. Strongly assumes IEEE. Faster than frexp. Obvious extension by bhudson.

int bithacks::log2_norm (  float  v  )  [inline]

Return the exponent of a normalized float. On a denorm, this returns the bias of -127. Strongly assumes IEEE. Faster than frexpf. Proper non-punning version of what's on the bithacks page.

---