void *__builtin_alloca(unsigned size)

Returns a pointer to a specified number of bytes of uninitialized local stack space.

float __builtin_fsqrt(float x)

Computes the non-negative square root of a floating point argument.

double __builtin_sqrt(double x)

Computes the non-negative square root of a double argument.

float __builtin_fabs(float x)

Computes the absolute value of a float argument.

double __builtin_dabs(double x)

Computes the absolute value of a double argument.

int __builtin_cast_f2i(float x)

Treats float as int.

float __builtin_cast_i2f(int x)

Treats int as float.

long long __builtin_cast_d2ll(double x)

Treats double as long long.

double __builtin_cast_ll2d(long long x)

Treats long long as double.

int __builtin_copy_dhi2i(double x)

Copies high part of double to int.

double __builtin_copy_i2dhi(int x)

Copies int to high part of double.

int __builtin_copy_dlo2i(double x)

Copies low part of double to int.

double __builtin_copy_i2dlo(int x, double y)

Copies int to low part of double.

<type> __high_multiply (<type>,<type> )

Multiplies two parameters as 32 (or 64) bit integers and returns the upper 32 (or 64) bits of a 64 (or 128) bit result. <type> can be signed or unsigned, int, long, or long long.