Re: using linux instead of osf

David Mosberger (
Tue, 26 Nov 1996 10:26:07 -0700 (MST)

> Well I think the GNU libm is so primitive that it must be updated, there's
> no single precision version of many of the functions so you always work in
> double precision whether you need it or not.

This is no longer true. Any ELF-based distribution comes with a
recent glibc and that library comes with fdlibm. The single-precision
operations are available as the function name with an `f' appended
(e.g., sinf() to get a `float' sin).

Not to say that the current fdlibm implementation is optimal for the
Alpha (far from that), but single-precision ops are definitively


To unsubscribe: send e-mail to with
'unsubscribe' as the subject.  Do not send it to

Feedback | Store | News | Support | Product Errata | About Us | Linux Info | Search | JumpWords
No Frames | Show Frames

Copyright © 1995-1997 Red Hat Software. Legal notices