mirror of
https://codeberg.org/ziglang/zig.git
synced 2025-12-06 05:44:20 +00:00
If you write an if expression in mem.doNotOptimizeAway like doNotOptimizeAway(if (ix < 0x00100000) x / 0x1p120 else x + 0x1p120);, FCSEL instruction is used on AArch64. FCSEL instruction selects one of the two registers according to the condition and copies its value. In this example, `x / 0x1p120` and `x + 0x1p120` are expressions that raise different floating-point exceptions. However, since both are actually evaluated before the FCSEL instruction, the exception not intended by the programmer may also be raised. To prevent FCSEL instruction from being used here, this commit splits doNotOptimizeAway in two.
203 lines
6.6 KiB
Zig
203 lines
6.6 KiB
Zig
//! Ported from musl, which is licensed under the MIT license:
|
|
//! https://git.musl-libc.org/cgit/musl/tree/COPYRIGHT
|
|
//!
|
|
//! https://git.musl-libc.org/cgit/musl/tree/src/math/sinf.c
|
|
//! https://git.musl-libc.org/cgit/musl/tree/src/math/sin.c
|
|
|
|
const std = @import("std");
|
|
const builtin = @import("builtin");
|
|
const arch = builtin.cpu.arch;
|
|
const math = std.math;
|
|
const mem = std.mem;
|
|
const expect = std.testing.expect;
|
|
const common = @import("common.zig");
|
|
|
|
const trig = @import("trig.zig");
|
|
const rem_pio2 = @import("rem_pio2.zig").rem_pio2;
|
|
const rem_pio2f = @import("rem_pio2f.zig").rem_pio2f;
|
|
|
|
pub const panic = common.panic;
|
|
|
|
comptime {
|
|
@export(&__sinh, .{ .name = "__sinh", .linkage = common.linkage, .visibility = common.visibility });
|
|
@export(&sinf, .{ .name = "sinf", .linkage = common.linkage, .visibility = common.visibility });
|
|
@export(&sin, .{ .name = "sin", .linkage = common.linkage, .visibility = common.visibility });
|
|
@export(&__sinx, .{ .name = "__sinx", .linkage = common.linkage, .visibility = common.visibility });
|
|
if (common.want_ppc_abi) {
|
|
@export(&sinq, .{ .name = "sinf128", .linkage = common.linkage, .visibility = common.visibility });
|
|
}
|
|
@export(&sinq, .{ .name = "sinq", .linkage = common.linkage, .visibility = common.visibility });
|
|
@export(&sinl, .{ .name = "sinl", .linkage = common.linkage, .visibility = common.visibility });
|
|
}
|
|
|
|
pub fn __sinh(x: f16) callconv(.c) f16 {
|
|
// TODO: more efficient implementation
|
|
return @floatCast(sinf(x));
|
|
}
|
|
|
|
pub fn sinf(x: f32) callconv(.c) f32 {
|
|
// Small multiples of pi/2 rounded to double precision.
|
|
const s1pio2: f64 = 1.0 * math.pi / 2.0; // 0x3FF921FB, 0x54442D18
|
|
const s2pio2: f64 = 2.0 * math.pi / 2.0; // 0x400921FB, 0x54442D18
|
|
const s3pio2: f64 = 3.0 * math.pi / 2.0; // 0x4012D97C, 0x7F3321D2
|
|
const s4pio2: f64 = 4.0 * math.pi / 2.0; // 0x401921FB, 0x54442D18
|
|
|
|
var ix: u32 = @bitCast(x);
|
|
const sign = ix >> 31 != 0;
|
|
ix &= 0x7fffffff;
|
|
|
|
if (ix <= 0x3f490fda) { // |x| ~<= pi/4
|
|
if (ix < 0x39800000) { // |x| < 2**-12
|
|
// raise inexact if x!=0 and underflow if subnormal
|
|
if (common.want_float_exceptions) {
|
|
if (ix < 0x00800000) {
|
|
mem.doNotOptimizeAway(x / 0x1p120);
|
|
} else {
|
|
mem.doNotOptimizeAway(x + 0x1p120);
|
|
}
|
|
}
|
|
return x;
|
|
}
|
|
return trig.__sindf(x);
|
|
}
|
|
if (ix <= 0x407b53d1) { // |x| ~<= 5*pi/4
|
|
if (ix <= 0x4016cbe3) { // |x| ~<= 3pi/4
|
|
if (sign) {
|
|
return -trig.__cosdf(x + s1pio2);
|
|
} else {
|
|
return trig.__cosdf(x - s1pio2);
|
|
}
|
|
}
|
|
return trig.__sindf(if (sign) -(x + s2pio2) else -(x - s2pio2));
|
|
}
|
|
if (ix <= 0x40e231d5) { // |x| ~<= 9*pi/4
|
|
if (ix <= 0x40afeddf) { // |x| ~<= 7*pi/4
|
|
if (sign) {
|
|
return trig.__cosdf(x + s3pio2);
|
|
} else {
|
|
return -trig.__cosdf(x - s3pio2);
|
|
}
|
|
}
|
|
return trig.__sindf(if (sign) x + s4pio2 else x - s4pio2);
|
|
}
|
|
|
|
// sin(Inf or NaN) is NaN
|
|
if (ix >= 0x7f800000) {
|
|
return x - x;
|
|
}
|
|
|
|
var y: f64 = undefined;
|
|
const n = rem_pio2f(x, &y);
|
|
return switch (n & 3) {
|
|
0 => trig.__sindf(y),
|
|
1 => trig.__cosdf(y),
|
|
2 => trig.__sindf(-y),
|
|
else => -trig.__cosdf(y),
|
|
};
|
|
}
|
|
|
|
pub fn sin(x: f64) callconv(.c) f64 {
|
|
var ix = @as(u64, @bitCast(x)) >> 32;
|
|
ix &= 0x7fffffff;
|
|
|
|
// |x| ~< pi/4
|
|
if (ix <= 0x3fe921fb) {
|
|
if (ix < 0x3e500000) { // |x| < 2**-26
|
|
// raise inexact if x != 0 and underflow if subnormal
|
|
if (common.want_float_exceptions) {
|
|
if (ix < 0x00100000) {
|
|
mem.doNotOptimizeAway(x / 0x1p120);
|
|
} else {
|
|
mem.doNotOptimizeAway(x + 0x1p120);
|
|
}
|
|
}
|
|
return x;
|
|
}
|
|
return trig.__sin(x, 0.0, 0);
|
|
}
|
|
|
|
// sin(Inf or NaN) is NaN
|
|
if (ix >= 0x7ff00000) {
|
|
return x - x;
|
|
}
|
|
|
|
var y: [2]f64 = undefined;
|
|
const n = rem_pio2(x, &y);
|
|
return switch (n & 3) {
|
|
0 => trig.__sin(y[0], y[1], 1),
|
|
1 => trig.__cos(y[0], y[1]),
|
|
2 => -trig.__sin(y[0], y[1], 1),
|
|
else => -trig.__cos(y[0], y[1]),
|
|
};
|
|
}
|
|
|
|
pub fn __sinx(x: f80) callconv(.c) f80 {
|
|
// TODO: more efficient implementation
|
|
return @floatCast(sinq(x));
|
|
}
|
|
|
|
pub fn sinq(x: f128) callconv(.c) f128 {
|
|
// TODO: more correct implementation
|
|
return sin(@floatCast(x));
|
|
}
|
|
|
|
pub fn sinl(x: c_longdouble) callconv(.c) c_longdouble {
|
|
switch (@typeInfo(c_longdouble).float.bits) {
|
|
16 => return __sinh(x),
|
|
32 => return sinf(x),
|
|
64 => return sin(x),
|
|
80 => return __sinx(x),
|
|
128 => return sinq(x),
|
|
else => @compileError("unreachable"),
|
|
}
|
|
}
|
|
|
|
test "sin32" {
|
|
const epsilon = 0.00001;
|
|
|
|
try expect(math.approxEqAbs(f32, sinf(0.0), 0.0, epsilon));
|
|
try expect(math.approxEqAbs(f32, sinf(0.2), 0.198669, epsilon));
|
|
try expect(math.approxEqAbs(f32, sinf(0.8923), 0.778517, epsilon));
|
|
try expect(math.approxEqAbs(f32, sinf(1.5), 0.997495, epsilon));
|
|
try expect(math.approxEqAbs(f32, sinf(-1.5), -0.997495, epsilon));
|
|
try expect(math.approxEqAbs(f32, sinf(37.45), -0.246544, epsilon));
|
|
try expect(math.approxEqAbs(f32, sinf(89.123), 0.916166, epsilon));
|
|
}
|
|
|
|
test "sin64" {
|
|
const epsilon = 0.000001;
|
|
|
|
try expect(math.approxEqAbs(f64, sin(0.0), 0.0, epsilon));
|
|
try expect(math.approxEqAbs(f64, sin(0.2), 0.198669, epsilon));
|
|
try expect(math.approxEqAbs(f64, sin(0.8923), 0.778517, epsilon));
|
|
try expect(math.approxEqAbs(f64, sin(1.5), 0.997495, epsilon));
|
|
try expect(math.approxEqAbs(f64, sin(-1.5), -0.997495, epsilon));
|
|
try expect(math.approxEqAbs(f64, sin(37.45), -0.246543, epsilon));
|
|
try expect(math.approxEqAbs(f64, sin(89.123), 0.916166, epsilon));
|
|
}
|
|
|
|
test "sin32.special" {
|
|
try expect(sinf(0.0) == 0.0);
|
|
try expect(sinf(-0.0) == -0.0);
|
|
try expect(math.isNan(sinf(math.inf(f32))));
|
|
try expect(math.isNan(sinf(-math.inf(f32))));
|
|
try expect(math.isNan(sinf(math.nan(f32))));
|
|
}
|
|
|
|
test "sin64.special" {
|
|
try expect(sin(0.0) == 0.0);
|
|
try expect(sin(-0.0) == -0.0);
|
|
try expect(math.isNan(sin(math.inf(f64))));
|
|
try expect(math.isNan(sin(-math.inf(f64))));
|
|
try expect(math.isNan(sin(math.nan(f64))));
|
|
}
|
|
|
|
test "sin32 #9901" {
|
|
const float: f32 = @bitCast(@as(u32, 0b11100011111111110000000000000000));
|
|
_ = sinf(float);
|
|
}
|
|
|
|
test "sin64 #9901" {
|
|
const float: f64 = @bitCast(@as(u64, 0b1111111101000001000000001111110111111111100000000000000000000001));
|
|
_ = sin(float);
|
|
}
|