const math.bias

21 uses

	math (current package)
		bits.go#L14: 	bias     = 1023
		floor.go#L91: 	if e < bias {
		floor.go#L94: 		if e == bias-1 {
		floor.go#L97: 	} else if e < bias+shift {
		floor.go#L103: 		e -= bias
		floor.go#L129: 	if e >= bias {
		floor.go#L135: 		e -= bias
		floor.go#L138: 	} else if e == bias-1 && bits&fracMask != 0 {
		fma.go#L116: 	pe := xe + ye - bias + 1
		fma.go#L158: 	if pe > 1022+bias || pe == 1022+bias && (m+1<<9)>>63 == 1 {
		frexp.go#L33: 	exp += int((x>>shift)&mask) - bias + 1
		frexp.go#L35: 	x |= (-1 + bias) << shift
		ldexp.go#L32: 	exp += int(x>>shift)&mask - bias
		ldexp.go#L48: 	x |= uint64(exp+bias) << shift
		logb.go#L49: 	return int((Float64bits(x)>>shift)&mask) - bias + exp
		modf.go#L33: 	e := uint(x>>shift)&mask - bias
		sqrt.go#L122: 	exp -= bias // unbias exponent
		sqrt.go#L147: 	ix = q>>1 + uint64(exp-1+bias)<<shift // significand + biased exponent
		trig_reduce.go#L37: 	exp := int(ix>>shift&mask) - bias - shift
		trig_reduce.go#L58: 	e := uint64(bias - (lz + 1))