Has anyone on the list seen this?  Is it better than what we have?
Would this solve the performance issues recently discussed?  These are
some of the questions this raises for me.  I am on my way to grab it.


>From dgh@validgh.com (David G. Hough at validgh)
>Newsgroups: comp.sys.sun.misc,comp.sys.ibm.pc.misc,comp.sys.mac.misc,comp.sys.misc,comp.sys.amiga.misc,comp.sys.dec,comp.sys.hp,comp.sys.sgi,alt.sys.sun
>Subject: freely distributable libm available
>Date: 18 Dec 93 19:06:13 GMT
>
>
>Dr. K-C Ng has developed a new version of libm that is the basis for
>the bundled /usr/lib/libm.so in Solaris 2.3 for SPARC and for future
>Solaris 2 releases for x86 and PowerPC.   It provides the standard
>functions necessary to pass the usual test suites.
>This new libm can be configured to handle exceptions in accordance with
>various language standards or in the spirit of IEEE 754.
>The C source code should be portable to any IEEE 754 system with minimal
>difficulty.
>
>fdlibm contains no additional functions
>beyond those required by various standards;
>the additional functions that used to be part of Sun libm
>will be distributed in a new library libsunmath.a that will
>be shipped with the unbundled SunPro compilers.
>
>In the interest of elevating the minimum acceptable level of quality in
>math libraries, SunPro has decided to make K-C's work freely distributable
>subject only the restriction that its origin be acknowledged.
>
>How to get fdlibm by email - as three email messages of 100KB each:
>
>   	mail netlib@research.att.com
>   	send all from fdlibm
> 
>How to get fdlibm by ftp:
>
>	ftp netlib.att.com
>   	cd netlib
>   	get fdlibm.tar
>
>Remember that anything free comes with no guarantee.  But
>you may send comments and bug reports to:
>       
>         fdlibm-comments@sunpro.eng.sun.com
>
>
>-- 
>
>David Hough
>dgh@validgh.com

------------------------------------------------------------------------------