executes the compute commands from an indirect command pipeline executes #"command_count" commands (or all if ~0u) starting at "command_offset" -> all

returns the extent of this vector (components multiplied with each other)

returns N if Nref.dot(this) < 0, else -N

returns an int vector containing 1 + the index of the least significant 1-bit of each component in this vector, or 0 if the component is 0

finds the n-th set bit (specified by "offset") in "value", starting at bit #"base", returns ~0u if the n-th set bit is not found NOTE: base must be >=

scalar passthrough

default empty map constructor

static function to convert a single precision float value to a half precision (stored as uint16_t)

static class

this is not supported (or in any way useful with just memory_order_relaxed)

from opencl_common: just forward to the context function

-> forward to lsb/msb functions

non-standard bit counting functions (don't use these directly, use math::func instead)

static class

fused-multiply-add copy of "(this * b_vec) + c_vec"