diff options
| author | Eric Botcazou <ebotcazou@libertysurf.fr> | 2003-09-22 07:46:51 +0200 |
|---|---|---|
| committer | Eric Botcazou <ebotcazou@gcc.gnu.org> | 2003-09-22 05:46:51 +0000 |
| commit | 9554f886547ff68f6ba5775f96ec14f59fb457ef (patch) | |
| tree | 6a49c65764b2ef13f9b1746cdbd9440a0778de38 | |
| parent | f31686a322301454c5ea5b997997aa7c84c57781 (diff) | |
| download | gcc-9554f886547ff68f6ba5775f96ec14f59fb457ef.zip gcc-9554f886547ff68f6ba5775f96ec14f59fb457ef.tar.gz gcc-9554f886547ff68f6ba5775f96ec14f59fb457ef.tar.bz2 | |
* real.c: Fix several nits in the head comment.
From-SVN: r71642
| -rw-r--r-- | gcc/ChangeLog | 4 | ||||
| -rw-r--r-- | gcc/real.c | 18 |
2 files changed, 13 insertions, 9 deletions
diff --git a/gcc/ChangeLog b/gcc/ChangeLog index 0b9c8e6..c171f9f 100644 --- a/gcc/ChangeLog +++ b/gcc/ChangeLog @@ -1,3 +1,7 @@ +2003-09-22 Eric Botcazou <ebotcazou@libertysurf.fr> + + * real.c: Fix several nits in the head comment. + 2003-09-21 Richard Henderson <rth@redhat.com> * tree.h c-aux-info.c, c-decl.c, c-parse.in, coverage.c, dbxout.c, @@ -2,7 +2,7 @@ Copyright (C) 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2002, 2003 Free Software Foundation, Inc. Contributed by Stephen L. Moshier (moshier@world.std.com). - Re-written by Richard Henderson <rth@redhat.com> + Re-written by Richard Henderson <rth@redhat.com> This file is part of GCC. @@ -31,7 +31,7 @@ #include "tm_p.h" /* The floating point model used internally is not exactly IEEE 754 - compliant, and close to the description in the ISO C standard, + compliant, and close to the description in the ISO C99 standard, section 5.2.4.2.2 Characteristics of floating types. Specifically @@ -49,21 +49,21 @@ significand is fractional. Normalized significands are in the range [0.5, 1.0). - A requirement of the model is that P be larger than than the - largest supported target floating-point type by at least 2 bits. - This gives us proper rounding when we truncate to the target type. - In addition, E must be large enough to hold the smallest supported - denormal number in a normalized form. + A requirement of the model is that P be larger than the largest + supported target floating-point type by at least 2 bits. This gives + us proper rounding when we truncate to the target type. In addition, + E must be large enough to hold the smallest supported denormal number + in a normalized form. Both of these requirements are easily satisfied. The largest target significand is 113 bits; we store at least 160. The smallest denormal number fits in 17 exponent bits; we store 29. Note that the decimal string conversion routines are sensitive to - rounding error. Since the raw arithmetic routines do not themselves + rounding errors. Since the raw arithmetic routines do not themselves have guard digits or rounding, the computation of 10**exp can accumulate more than a few digits of error. The previous incarnation - of real.c successfully used a 144 bit fraction; given the current + of real.c successfully used a 144-bit fraction; given the current layout of REAL_VALUE_TYPE we're forced to expand to at least 160 bits. Target floating point models that use base 16 instead of base 2 |