add f16 type

Add support for half-precision floating point operations.

Introduce `__extendhfsf2` and `__truncsfhf2` in std/special/compiler_rt.

Add `__gnu_h2f_ieee` and `__gnu_f2h_ieee` as aliases that are used in
Windows builds.

The logic in std/special/compiler_rt/extendXfYf2.zig has been reworked
and can now operate on 16 bits floating point types.

`extendXfYf2()` and `truncXfYf2()` are marked `inline` to work around
a not entirely understood stack alignment issue on Windows when calling
the f16 versions of the builtins.

closes #1122

1 files changed, 8 insertions, 0 deletions

diff --git a/src/bigfloat.cpp b/src/bigfloat.cpp
index dcb6db61db..cc442fa3b7 100644
--- a/src/bigfloat.cpp
+++ b/src/bigfloat.cpp
@@ -18,6 +18,10 @@ void bigfloat_init_128(BigFloat *dest, float128_t x) {
     dest->value = x;
 }
 
+void bigfloat_init_16(BigFloat *dest, float16_t x) {
+    f16_to_f128M(x, &dest->value);
+}
+
 void bigfloat_init_32(BigFloat *dest, float x) {
     float32_t f32_val;
     memcpy(&f32_val, &x, sizeof(float));
@@ -146,6 +150,10 @@ Cmp bigfloat_cmp(const BigFloat *op1, const BigFloat *op2) {
     }
 }
 
+float16_t bigfloat_to_f16(const BigFloat *bigfloat) {
+    return f128M_to_f16(&bigfloat->value);
+}
+
 float bigfloat_to_f32(const BigFloat *bigfloat) {
     float32_t f32_value = f128M_to_f32(&bigfloat->value);
     float result;