[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Bug-gnulib] [RFC] Module: long double math functions
From: |
Paolo Bonzini |
Subject: |
Re: [Bug-gnulib] [RFC] Module: long double math functions |
Date: |
Tue, 11 Feb 2003 09:16:38 +0100 |
User-agent: |
KMail/1.4.2 |
> sizeof (long double) > sizeof (double)
Not strictly necessary. I.e. no infinite loops if sizeof (long double) ==
sizeof (double), only wasted time.
> precisionof (long double) <= 2 * precisionof (double)
Not strictly necessary. Results will not be precise to 1 ulp if the precision
of long doubles is greater than 113 bits, though (are there such monsters?).
> IEEE format
No. That's the main change I made WRT glibc's code.
> sizeof (long double) == 12 or == 16 ?
No.
> If you have precision problems, you can look at the algorithms of
> Keith Briggs' doubledouble package (a.k.a. xdouble in LIDIA), see
> http://members.lycos.co.uk/keithmbriggs/doubledouble.html
> - his algorithms provide 106 bits of precision.
These also do, though they of course rely on the compiler being able to
convert decimal constants properly.
Paolo