|
|
@@ -33,13 +33,6 @@ foreign _ {
|
|
|
@(link_name="llvm.fmuladd.f64")
|
|
|
fmuladd_f64 :: proc(a, b, c: f64) -> f64 ---
|
|
|
|
|
|
- @(link_name="llvm.log.f16")
|
|
|
- ln_f16 :: proc(x: f16) -> f16 ---
|
|
|
- @(link_name="llvm.log.f32")
|
|
|
- ln_f32 :: proc(x: f32) -> f32 ---
|
|
|
- @(link_name="llvm.log.f64")
|
|
|
- ln_f64 :: proc(x: f64) -> f64 ---
|
|
|
-
|
|
|
@(link_name="llvm.exp.f16")
|
|
|
exp_f16 :: proc(x: f16) -> f16 ---
|
|
|
@(link_name="llvm.exp.f32")
|
|
|
@@ -57,3 +50,120 @@ sqrt_f32 :: proc "contextless" (x: f32) -> f32 {
|
|
|
sqrt_f64 :: proc "contextless" (x: f64) -> f64 {
|
|
|
return intrinsics.sqrt(x)
|
|
|
}
|
|
|
+
|
|
|
+
|
|
|
+
|
|
|
+ln_f64 :: proc "contextless" (x: f64) -> f64 {
|
|
|
+ // The original C code, the long comment, and the constants
|
|
|
+ // below are from FreeBSD's /usr/src/lib/msun/src/e_log.c
|
|
|
+ // and came with this notice.
|
|
|
+ //
|
|
|
+ // ====================================================
|
|
|
+ // Copyright (C) 1993 by Sun Microsystems, Inc. All rights reserved.
|
|
|
+ //
|
|
|
+ // Developed at SunPro, a Sun Microsystems, Inc. business.
|
|
|
+ // Permission to use, copy, modify, and distribute this
|
|
|
+ // software is freely granted, provided that this notice
|
|
|
+ // is preserved.
|
|
|
+ // ====================================================
|
|
|
+ //
|
|
|
+ // __ieee754_log(x)
|
|
|
+ // Return the logarithm of x
|
|
|
+ //
|
|
|
+ // Method :
|
|
|
+ // 1. Argument Reduction: find k and f such that
|
|
|
+ // x = 2**k * (1+f),
|
|
|
+ // where sqrt(2)/2 < 1+f < sqrt(2) .
|
|
|
+ //
|
|
|
+ // 2. Approximation of log(1+f).
|
|
|
+ // Let s = f/(2+f) ; based on log(1+f) = log(1+s) - log(1-s)
|
|
|
+ // = 2s + 2/3 s**3 + 2/5 s**5 + .....,
|
|
|
+ // = 2s + s*R
|
|
|
+ // We use a special Reme algorithm on [0,0.1716] to generate
|
|
|
+ // a polynomial of degree 14 to approximate R. The maximum error
|
|
|
+ // of this polynomial approximation is bounded by 2**-58.45. In
|
|
|
+ // other words,
|
|
|
+ // 2 4 6 8 10 12 14
|
|
|
+ // R(z) ~ L1*s +L2*s +L3*s +L4*s +L5*s +L6*s +L7*s
|
|
|
+ // (the values of L1 to L7 are listed in the program) and
|
|
|
+ // | 2 14 | -58.45
|
|
|
+ // | L1*s +...+L7*s - R(z) | <= 2
|
|
|
+ // | |
|
|
|
+ // Note that 2s = f - s*f = f - hfsq + s*hfsq, where hfsq = f*f/2.
|
|
|
+ // In order to guarantee error in log below 1ulp, we compute log by
|
|
|
+ // log(1+f) = f - s*(f - R) (if f is not too large)
|
|
|
+ // log(1+f) = f - (hfsq - s*(hfsq+R)). (better accuracy)
|
|
|
+ //
|
|
|
+ // 3. Finally, log(x) = k*Ln2 + log(1+f).
|
|
|
+ // = k*Ln2_hi+(f-(hfsq-(s*(hfsq+R)+k*Ln2_lo)))
|
|
|
+ // Here Ln2 is split into two floating point number:
|
|
|
+ // Ln2_hi + Ln2_lo,
|
|
|
+ // where n*Ln2_hi is always exact for |n| < 2000.
|
|
|
+ //
|
|
|
+ // Special cases:
|
|
|
+ // log(x) is NaN with signal if x < 0 (including -INF) ;
|
|
|
+ // log(+INF) is +INF; log(0) is -INF with signal;
|
|
|
+ // log(NaN) is that NaN with no signal.
|
|
|
+ //
|
|
|
+ // Accuracy:
|
|
|
+ // according to an error analysis, the error is always less than
|
|
|
+ // 1 ulp (unit in the last place).
|
|
|
+ //
|
|
|
+ // Constants:
|
|
|
+ // The hexadecimal values are the intended ones for the following
|
|
|
+ // constants. The decimal values may be used, provided that the
|
|
|
+ // compiler will convert from decimal to binary accurately enough
|
|
|
+ // to produce the hexadecimal values shown.
|
|
|
+
|
|
|
+ LN2_HI :: 0h3fe62e42_fee00000 // 6.93147180369123816490e-01
|
|
|
+ LN2_LO :: 0h3dea39ef_35793c76 // 1.90821492927058770002e-10
|
|
|
+ L1 :: 0h3fe55555_55555593 // 6.666666666666735130e-01
|
|
|
+ L2 :: 0h3fd99999_9997fa04 // 3.999999999940941908e-01
|
|
|
+ L3 :: 0h3fd24924_94229359 // 2.857142874366239149e-01
|
|
|
+ L4 :: 0h3fcc71c5_1d8e78af // 2.222219843214978396e-01
|
|
|
+ L5 :: 0h3fc74664_96cb03de // 1.818357216161805012e-01
|
|
|
+ L6 :: 0h3fc39a09_d078c69f // 1.531383769920937332e-01
|
|
|
+ L7 :: 0h3fc2f112_df3e5244 // 1.479819860511658591e-01
|
|
|
+
|
|
|
+ switch {
|
|
|
+ case is_nan(x) || is_inf(x, 1):
|
|
|
+ return x
|
|
|
+ case x < 0:
|
|
|
+ return nan_f64()
|
|
|
+ case x == 0:
|
|
|
+ return inf_f64(-1)
|
|
|
+ }
|
|
|
+
|
|
|
+ // reduce
|
|
|
+ f1, ki := frexp(x)
|
|
|
+ if f1 < SQRT_TWO/2 {
|
|
|
+ f1 *= 2
|
|
|
+ ki -= 1
|
|
|
+ }
|
|
|
+ f := f1 - 1
|
|
|
+ k := f64(ki)
|
|
|
+
|
|
|
+ // compute
|
|
|
+ s := f / (2 + f)
|
|
|
+ s2 := s * s
|
|
|
+ s4 := s2 * s2
|
|
|
+ t1 := s2 * (L1 + s4*(L3+s4*(L5+s4*L7)))
|
|
|
+ t2 := s4 * (L2 + s4*(L4+s4*L6))
|
|
|
+ R := t1 + t2
|
|
|
+ hfsq := 0.5 * f * f
|
|
|
+ return k*Ln2Hi_ - ((hfsq - (s*(hfsq+R) + k*LN2_LO)) - f)
|
|
|
+}
|
|
|
+
|
|
|
+ln_f16 :: proc "contextless" (x: f16) -> f16 { return #force_inline f16(ln_f64(f64(x))) }
|
|
|
+ln_f32 :: proc "contextless" (x: f32) -> f32 { return #force_inline f32(ln_f64(f64(x))) }
|
|
|
+ln_f16le :: proc "contextless" (x: f16le) -> f16le { return #force_inline f16le(ln_f64(f64(x))) }
|
|
|
+ln_f16be :: proc "contextless" (x: f16be) -> f16be { return #force_inline f16be(ln_f64(f64(x))) }
|
|
|
+ln_f32le :: proc "contextless" (x: f32le) -> f32le { return #force_inline f32le(ln_f64(f64(x))) }
|
|
|
+ln_f32be :: proc "contextless" (x: f32be) -> f32be { return #force_inline f32be(ln_f64(f64(x))) }
|
|
|
+ln_f64le :: proc "contextless" (x: f64le) -> f64le { return #force_inline f64le(ln_f64(f64(x))) }
|
|
|
+ln_f64be :: proc "contextless" (x: f64be) -> f64be { return #force_inline f64be(ln_f64(f64(x))) }
|
|
|
+ln :: proc{
|
|
|
+ ln_f16, ln_f16le, ln_f16be,
|
|
|
+ ln_f32, ln_f32le, ln_f32be,
|
|
|
+ ln_f64, ln_f64le, ln_f64be,
|
|
|
+}
|
gingerBill