zig/lib/compiler_rt/divtf3.zig
2022-06-17 16:38:59 -07:00

251 lines
9.7 KiB
Zig

const std = @import("std");
const builtin = @import("builtin");
const common = @import("common.zig");
const normalize = common.normalize;
const wideMultiply = common.wideMultiply;
pub const panic = common.panic;
comptime {
if (common.want_ppc_abi) {
@export(__divkf3, .{ .name = "__divkf3", .linkage = common.linkage });
} else if (common.want_sparc_abi) {
@export(_Qp_div, .{ .name = "_Qp_div", .linkage = common.linkage });
} else {
@export(__divtf3, .{ .name = "__divtf3", .linkage = common.linkage });
}
}
pub fn __divtf3(a: f128, b: f128) callconv(.C) f128 {
return div(a, b);
}
fn __divkf3(a: f128, b: f128) callconv(.C) f128 {
return div(a, b);
}
fn _Qp_div(c: *f128, a: *const f128, b: *const f128) callconv(.C) void {
c.* = div(a.*, b.*);
}
inline fn div(a: f128, b: f128) f128 {
const Z = std.meta.Int(.unsigned, 128);
const significandBits = std.math.floatMantissaBits(f128);
const exponentBits = std.math.floatExponentBits(f128);
const signBit = (@as(Z, 1) << (significandBits + exponentBits));
const maxExponent = ((1 << exponentBits) - 1);
const exponentBias = (maxExponent >> 1);
const implicitBit = (@as(Z, 1) << significandBits);
const quietBit = implicitBit >> 1;
const significandMask = implicitBit - 1;
const absMask = signBit - 1;
const exponentMask = absMask ^ significandMask;
const qnanRep = exponentMask | quietBit;
const infRep = @bitCast(Z, std.math.inf(f128));
const aExponent = @truncate(u32, (@bitCast(Z, a) >> significandBits) & maxExponent);
const bExponent = @truncate(u32, (@bitCast(Z, b) >> significandBits) & maxExponent);
const quotientSign: Z = (@bitCast(Z, a) ^ @bitCast(Z, b)) & signBit;
var aSignificand: Z = @bitCast(Z, a) & significandMask;
var bSignificand: Z = @bitCast(Z, b) & significandMask;
var scale: i32 = 0;
// Detect if a or b is zero, denormal, infinity, or NaN.
if (aExponent -% 1 >= maxExponent - 1 or bExponent -% 1 >= maxExponent - 1) {
const aAbs: Z = @bitCast(Z, a) & absMask;
const bAbs: Z = @bitCast(Z, b) & absMask;
// NaN / anything = qNaN
if (aAbs > infRep) return @bitCast(f128, @bitCast(Z, a) | quietBit);
// anything / NaN = qNaN
if (bAbs > infRep) return @bitCast(f128, @bitCast(Z, b) | quietBit);
if (aAbs == infRep) {
// infinity / infinity = NaN
if (bAbs == infRep) {
return @bitCast(f128, qnanRep);
}
// infinity / anything else = +/- infinity
else {
return @bitCast(f128, aAbs | quotientSign);
}
}
// anything else / infinity = +/- 0
if (bAbs == infRep) return @bitCast(f128, quotientSign);
if (aAbs == 0) {
// zero / zero = NaN
if (bAbs == 0) {
return @bitCast(f128, qnanRep);
}
// zero / anything else = +/- zero
else {
return @bitCast(f128, quotientSign);
}
}
// anything else / zero = +/- infinity
if (bAbs == 0) return @bitCast(f128, infRep | quotientSign);
// one or both of a or b is denormal, the other (if applicable) is a
// normal number. Renormalize one or both of a and b, and set scale to
// include the necessary exponent adjustment.
if (aAbs < implicitBit) scale +%= normalize(f128, &aSignificand);
if (bAbs < implicitBit) scale -%= normalize(f128, &bSignificand);
}
// Set the implicit significand bit. If we fell through from the
// denormal path it was already set by normalize( ), but setting it twice
// won't hurt anything.
aSignificand |= implicitBit;
bSignificand |= implicitBit;
var quotientExponent: i32 = @bitCast(i32, aExponent -% bExponent) +% scale;
// Align the significand of b as a Q63 fixed-point number in the range
// [1, 2.0) and get a Q64 approximate reciprocal using a small minimax
// polynomial approximation: reciprocal = 3/4 + 1/sqrt(2) - b/2. This
// is accurate to about 3.5 binary digits.
const q63b = @truncate(u64, bSignificand >> 49);
var recip64 = @as(u64, 0x7504f333F9DE6484) -% q63b;
// 0x7504f333F9DE6484 / 2^64 + 1 = 3/4 + 1/sqrt(2)
// Now refine the reciprocal estimate using a Newton-Raphson iteration:
//
// x1 = x0 * (2 - x0 * b)
//
// This doubles the number of correct binary digits in the approximation
// with each iteration.
var correction64: u64 = undefined;
correction64 = @truncate(u64, ~(@as(u128, recip64) *% q63b >> 64) +% 1);
recip64 = @truncate(u64, @as(u128, recip64) *% correction64 >> 63);
correction64 = @truncate(u64, ~(@as(u128, recip64) *% q63b >> 64) +% 1);
recip64 = @truncate(u64, @as(u128, recip64) *% correction64 >> 63);
correction64 = @truncate(u64, ~(@as(u128, recip64) *% q63b >> 64) +% 1);
recip64 = @truncate(u64, @as(u128, recip64) *% correction64 >> 63);
correction64 = @truncate(u64, ~(@as(u128, recip64) *% q63b >> 64) +% 1);
recip64 = @truncate(u64, @as(u128, recip64) *% correction64 >> 63);
correction64 = @truncate(u64, ~(@as(u128, recip64) *% q63b >> 64) +% 1);
recip64 = @truncate(u64, @as(u128, recip64) *% correction64 >> 63);
// The reciprocal may have overflowed to zero if the upper half of b is
// exactly 1.0. This would sabatoge the full-width final stage of the
// computation that follows, so we adjust the reciprocal down by one bit.
recip64 -%= 1;
// We need to perform one more iteration to get us to 112 binary digits;
// The last iteration needs to happen with extra precision.
const q127blo: u64 = @truncate(u64, bSignificand << 15);
var correction: u128 = undefined;
var reciprocal: u128 = undefined;
// NOTE: This operation is equivalent to __multi3, which is not implemented
// in some architechure
var r64q63: u128 = undefined;
var r64q127: u128 = undefined;
var r64cH: u128 = undefined;
var r64cL: u128 = undefined;
var dummy: u128 = undefined;
wideMultiply(u128, recip64, q63b, &dummy, &r64q63);
wideMultiply(u128, recip64, q127blo, &dummy, &r64q127);
correction = -%(r64q63 + (r64q127 >> 64));
const cHi = @truncate(u64, correction >> 64);
const cLo = @truncate(u64, correction);
wideMultiply(u128, recip64, cHi, &dummy, &r64cH);
wideMultiply(u128, recip64, cLo, &dummy, &r64cL);
reciprocal = r64cH + (r64cL >> 64);
// Adjust the final 128-bit reciprocal estimate downward to ensure that it
// is strictly smaller than the infinitely precise exact reciprocal. Because
// the computation of the Newton-Raphson step is truncating at every step,
// this adjustment is small; most of the work is already done.
reciprocal -%= 2;
// The numerical reciprocal is accurate to within 2^-112, lies in the
// interval [0.5, 1.0), and is strictly smaller than the true reciprocal
// of b. Multiplying a by this reciprocal thus gives a numerical q = a/b
// in Q127 with the following properties:
//
// 1. q < a/b
// 2. q is in the interval [0.5, 2.0)
// 3. The error in q is bounded away from 2^-113 (actually, we have a
// couple of bits to spare, but this is all we need).
// We need a 128 x 128 multiply high to compute q.
var quotient: u128 = undefined;
var quotientLo: u128 = undefined;
wideMultiply(u128, aSignificand << 2, reciprocal, &quotient, &quotientLo);
// Two cases: quotient is in [0.5, 1.0) or quotient is in [1.0, 2.0).
// In either case, we are going to compute a residual of the form
//
// r = a - q*b
//
// We know from the construction of q that r satisfies:
//
// 0 <= r < ulp(q)*b
//
// If r is greater than 1/2 ulp(q)*b, then q rounds up. Otherwise, we
// already have the correct result. The exact halfway case cannot occur.
// We also take this time to right shift quotient if it falls in the [1,2)
// range and adjust the exponent accordingly.
var residual: u128 = undefined;
var qb: u128 = undefined;
if (quotient < (implicitBit << 1)) {
wideMultiply(u128, quotient, bSignificand, &dummy, &qb);
residual = (aSignificand << 113) -% qb;
quotientExponent -%= 1;
} else {
quotient >>= 1;
wideMultiply(u128, quotient, bSignificand, &dummy, &qb);
residual = (aSignificand << 112) -% qb;
}
const writtenExponent = quotientExponent +% exponentBias;
if (writtenExponent >= maxExponent) {
// If we have overflowed the exponent, return infinity.
return @bitCast(f128, infRep | quotientSign);
} else if (writtenExponent < 1) {
if (writtenExponent == 0) {
// Check whether the rounded result is normal.
const round = @boolToInt((residual << 1) > bSignificand);
// Clear the implicit bit.
var absResult = quotient & significandMask;
// Round.
absResult += round;
if ((absResult & ~significandMask) > 0) {
// The rounded result is normal; return it.
return @bitCast(f128, absResult | quotientSign);
}
}
// Flush denormals to zero. In the future, it would be nice to add
// code to round them correctly.
return @bitCast(f128, quotientSign);
} else {
const round = @boolToInt((residual << 1) >= bSignificand);
// Clear the implicit bit
var absResult = quotient & significandMask;
// Insert the exponent
absResult |= @intCast(Z, writtenExponent) << significandBits;
// Round
absResult +%= round;
// Insert the sign and return
return @bitCast(f128, absResult | quotientSign);
}
}
test {
_ = @import("divtf3_test.zig");
}