213 lines
5.2 KiB
Zig
213 lines
5.2 KiB
Zig
// Special Cases:
|
|
//
|
|
// - exp(+inf) = +inf
|
|
// - exp(nan) = nan
|
|
|
|
const std = @import("../index.zig");
|
|
const math = std.math;
|
|
const assert = std.debug.assert;
|
|
const builtin = @import("builtin");
|
|
|
|
pub fn exp(x: var) @typeOf(x) {
|
|
const T = @typeOf(x);
|
|
return switch (T) {
|
|
f32 => exp32(x),
|
|
f64 => exp64(x),
|
|
else => @compileError("exp not implemented for " ++ @typeName(T)),
|
|
};
|
|
}
|
|
|
|
fn exp32(x_: f32) f32 {
|
|
const half = []f32{ 0.5, -0.5 };
|
|
const ln2hi = 6.9314575195e-1;
|
|
const ln2lo = 1.4286067653e-6;
|
|
const invln2 = 1.4426950216e+0;
|
|
const P1 = 1.6666625440e-1;
|
|
const P2 = -2.7667332906e-3;
|
|
|
|
var x = x_;
|
|
var hx = @bitCast(u32, x);
|
|
const sign = @intCast(i32, hx >> 31);
|
|
hx &= 0x7FFFFFFF;
|
|
|
|
if (math.isNan(x)) {
|
|
return x;
|
|
}
|
|
|
|
// |x| >= -87.33655 or nan
|
|
if (hx >= 0x42AEAC50) {
|
|
// nan
|
|
if (hx > 0x7F800000) {
|
|
return x;
|
|
}
|
|
// x >= 88.722839
|
|
if (hx >= 0x42b17218 and sign == 0) {
|
|
return x * 0x1.0p127;
|
|
}
|
|
if (sign != 0) {
|
|
math.forceEval(-0x1.0p-149 / x); // overflow
|
|
// x <= -103.972084
|
|
if (hx >= 0x42CFF1B5) {
|
|
return 0;
|
|
}
|
|
}
|
|
}
|
|
|
|
var k: i32 = undefined;
|
|
var hi: f32 = undefined;
|
|
var lo: f32 = undefined;
|
|
|
|
// |x| > 0.5 * ln2
|
|
if (hx > 0x3EB17218) {
|
|
// |x| > 1.5 * ln2
|
|
if (hx > 0x3F851592) {
|
|
k = @floatToInt(i32, invln2 * x + half[@intCast(usize, sign)]);
|
|
} else {
|
|
k = 1 - sign - sign;
|
|
}
|
|
|
|
const fk = @intToFloat(f32, k);
|
|
hi = x - fk * ln2hi;
|
|
lo = fk * ln2lo;
|
|
x = hi - lo;
|
|
}
|
|
// |x| > 2^(-14)
|
|
else if (hx > 0x39000000) {
|
|
k = 0;
|
|
hi = x;
|
|
lo = 0;
|
|
} else {
|
|
math.forceEval(0x1.0p127 + x); // inexact
|
|
return 1 + x;
|
|
}
|
|
|
|
const xx = x * x;
|
|
const c = x - xx * (P1 + xx * P2);
|
|
const y = 1 + (x * c / (2 - c) - lo + hi);
|
|
|
|
if (k == 0) {
|
|
return y;
|
|
} else {
|
|
return math.scalbn(y, k);
|
|
}
|
|
}
|
|
|
|
fn exp64(x_: f64) f64 {
|
|
const half = []const f64{ 0.5, -0.5 };
|
|
const ln2hi: f64 = 6.93147180369123816490e-01;
|
|
const ln2lo: f64 = 1.90821492927058770002e-10;
|
|
const invln2: f64 = 1.44269504088896338700e+00;
|
|
const P1: f64 = 1.66666666666666019037e-01;
|
|
const P2: f64 = -2.77777777770155933842e-03;
|
|
const P3: f64 = 6.61375632143793436117e-05;
|
|
const P4: f64 = -1.65339022054652515390e-06;
|
|
const P5: f64 = 4.13813679705723846039e-08;
|
|
|
|
var x = x_;
|
|
var ux = @bitCast(u64, x);
|
|
var hx = ux >> 32;
|
|
const sign = @intCast(i32, hx >> 31);
|
|
hx &= 0x7FFFFFFF;
|
|
|
|
if (math.isNan(x)) {
|
|
return x;
|
|
}
|
|
|
|
// |x| >= 708.39 or nan
|
|
if (hx >= 0x4086232B) {
|
|
// nan
|
|
if (hx > 0x7FF00000) {
|
|
return x;
|
|
}
|
|
if (x > 709.782712893383973096) {
|
|
// overflow if x != inf
|
|
if (!math.isInf(x)) {
|
|
math.raiseOverflow();
|
|
}
|
|
return math.inf(f64);
|
|
}
|
|
if (x < -708.39641853226410622) {
|
|
// underflow if x != -inf
|
|
// math.forceEval(f32(-0x1.0p-149 / x));
|
|
if (x < -745.13321910194110842) {
|
|
return 0;
|
|
}
|
|
}
|
|
}
|
|
|
|
// argument reduction
|
|
var k: i32 = undefined;
|
|
var hi: f64 = undefined;
|
|
var lo: f64 = undefined;
|
|
|
|
// |x| > 0.5 * ln2
|
|
if (hx > 0x3EB17218) {
|
|
// |x| >= 1.5 * ln2
|
|
if (hx > 0x3FF0A2B2) {
|
|
k = @floatToInt(i32, invln2 * x + half[@intCast(usize, sign)]);
|
|
} else {
|
|
k = 1 - sign - sign;
|
|
}
|
|
|
|
const dk = @intToFloat(f64, k);
|
|
hi = x - dk * ln2hi;
|
|
lo = dk * ln2lo;
|
|
x = hi - lo;
|
|
}
|
|
// |x| > 2^(-28)
|
|
else if (hx > 0x3E300000) {
|
|
k = 0;
|
|
hi = x;
|
|
lo = 0;
|
|
} else {
|
|
// inexact if x != 0
|
|
// math.forceEval(0x1.0p1023 + x);
|
|
return 1 + x;
|
|
}
|
|
|
|
const xx = x * x;
|
|
const c = x - xx * (P1 + xx * (P2 + xx * (P3 + xx * (P4 + xx * P5))));
|
|
const y = 1 + (x * c / (2 - c) - lo + hi);
|
|
|
|
if (k == 0) {
|
|
return y;
|
|
} else {
|
|
return math.scalbn(y, k);
|
|
}
|
|
}
|
|
|
|
test "math.exp" {
|
|
assert(exp(f32(0.0)) == exp32(0.0));
|
|
assert(exp(f64(0.0)) == exp64(0.0));
|
|
}
|
|
|
|
test "math.exp32" {
|
|
const epsilon = 0.000001;
|
|
|
|
assert(exp32(0.0) == 1.0);
|
|
assert(math.approxEq(f32, exp32(0.0), 1.0, epsilon));
|
|
assert(math.approxEq(f32, exp32(0.2), 1.221403, epsilon));
|
|
assert(math.approxEq(f32, exp32(0.8923), 2.440737, epsilon));
|
|
assert(math.approxEq(f32, exp32(1.5), 4.481689, epsilon));
|
|
}
|
|
|
|
test "math.exp64" {
|
|
const epsilon = 0.000001;
|
|
|
|
assert(exp64(0.0) == 1.0);
|
|
assert(math.approxEq(f64, exp64(0.0), 1.0, epsilon));
|
|
assert(math.approxEq(f64, exp64(0.2), 1.221403, epsilon));
|
|
assert(math.approxEq(f64, exp64(0.8923), 2.440737, epsilon));
|
|
assert(math.approxEq(f64, exp64(1.5), 4.481689, epsilon));
|
|
}
|
|
|
|
test "math.exp32.special" {
|
|
assert(math.isPositiveInf(exp32(math.inf(f32))));
|
|
assert(math.isNan(exp32(math.nan(f32))));
|
|
}
|
|
|
|
test "math.exp64.special" {
|
|
assert(math.isPositiveInf(exp64(math.inf(f64))));
|
|
assert(math.isNan(exp64(math.nan(f64))));
|
|
}
|