strtod and dtoa correct rounding and fast
C Python
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
README.md
dtoa-aux.c
dtoa-fast.c
dtoa-fast.txt
dtoa-ifmt1.c
dtoa-ifmt2.c
dtoa-lite.c
dtoa-mode.c
dtoa.c
license.txt
mapm-src.7z
max_ulp.py
max_ulp.txt
strtod-aux.c
strtod-fast.c
strtod-fast.txt
strtod-gmp.c
strtod-lite.c
strtod-mode.c
strtod-mpfr.c
test.c

README.md

dtoa-fast

strtod and dtoa correct rounding FAST (only require 53 bits precision)

--> can go even FASTER with C-flag -D FAST_BF96

Both uses normalized 96-bits big float for calculations

If estimate is far from halfway, no need for arbitrary precision math

If estimate is guaranteed halfway, no need for arbitrary precision math

Remaining hard cases uses arbitrary precision math library to break ties.

X math libraries to use with -D USE_MAPM=X
0 use GMP for hard cases (note: GMP had GPL license)
1 link with MAPM v5.0 libmapm.a (DEFAULT setting)
2 same as 1, but use grade-school multiply only (no FFT)
3 same as 2, but #include as 1 BIG file (no need for libmapm.a)
96-bit float nice properties Eliminated bugs
no concept of normals/denormals no boundary crossing bug
no need for correction loops no infinite loops bug
no need to set precision mode no double rounding bug
Files Comments
dtoa-aux.c common routine for dtoa-fast.c and dtoa-mode.c
dtoa-fast.c double to string, use FE_TONEAREST mode
dtoa-fast.txt algorithm used for dtoa-fast.c and dtoa-mode.c
dtoa-ifmt1.c in-place format dtoa_fast() result, mode allowed = [regREG]
dtoa-ifmt2.c same as above, but returned shortest string, e.g. "1e9"
dtoa-lite.c dtoa-fast without bignum library, accurary 99.95%
dtoa-mode.c double to string, honors rounding mode
dtoa.c         David Gay's dtoa.c
mapm-src.7z     MAPM C Library v5.0
max_ulp.py simulation shows max ulp error at most 95 ULP
max_ulp.txt PROVED max ulp error is at most 95 ULP
strtod-aux.c use MAPM to break ties for hard cases
strtod-fast.c string to double, use FE_TONEAREST mode
strtod-fast.txt algorithm used for strtod-fast.c and strtod-mode.c
strtod-lite.c strtod-fast without bignum library, accuracy almost 100%
strtod-gmp.c use GMP to break ties for hard cases
strtod-mode.c string to double, honors rounding mode
strtod-mpfr.c strtod implementation using MPFR
test.c strtod, dtoa, and round-trip tests of my routines