Re: Machine epsilon: conclusion
- From: Keith Thompson <kst-u@xxxxxxx>
- Date: Sat, 30 Jun 2007 13:23:06 -0700
jacob navia <jacob@xxxxxxxxxxxxxxxx> writes:
Following the discussion about the machine epsilon, I have modified the
text as follows. I thank all people that contributed. It was
interesting, and I surely learned stuff I did not consider before.
It's good to see you responding well to criticism. This is a great
improvement over your recent "Jacob is always wrong" whining. Please
keep it up.
--------------------------------------------------------------
The machine epsilon
-------------------
The standard defines for each floating point representation (float,
double, long double) the difference between 1 and the least value
greater than 1 that is representable in the given floating point
type. In IEEE754 representation this number has an exponent value of
the bias, and a fraction of 1. In hexadecimal notation this would be
0x1p-52
Which representation does this apply to? There are several IEEE 754
representations (and I'd definitely put a space between IEEE and 754).
You should mention that the current name for that standard is IEC
60559; that's how the C standard refers to it. See C99 F.1p1.
You should also mention explicitly that the C standard does not
require IEC 60559 conformance; it merely provides optional support for
those implementations that do conform to it.
[...]
These definitions (except the qfloat part) are part of the C99 ANSI
standard. For the standard types (float, double and long double) they
should always exist in other compilers.
Why do you refer to the ISO C99 standard as "ANSI"?
[snip]
--
Keith Thompson (The_Other_Keith) kst-u@xxxxxxx <http://www.ghoti.net/~kst>
San Diego Supercomputer Center <*> <http://users.sdsc.edu/~kst>
"We must do something. This is something. Therefore, we must do this."
-- Antony Jay and Jonathan Lynn, "Yes Minister"
.
- References:
- Machine epsilon: conclusion
- From: jacob navia
- Machine epsilon: conclusion
- Prev by Date: Re: Machine epsilon: conclusion
- Next by Date: Re: reply the answer
- Previous by thread: Re: Machine epsilon: conclusion
- Next by thread: segmentation fault
- Index(es):
Relevant Pages
|
