Path: utzoo!utgpu!watmath!att!pacbell!ames!lll-winken!uunet!ibmpa!lmb From: lmb@ibmpa.UUCP (Larry Breed) Newsgroups: comp.sys.ibm.pc.rt Subject: Re: AFPA performance Summary: hf77 generates direct AFPA code Message-ID: <1678@ibmpa.UUCP> Date: 9 Aug 89 04:05:45 GMT References: <12926@bloom-beacon.MIT.EDU> <263@archet.UUCP> Reply-To: lmb@ibmpa.UUCP (Larry Breed) Organization: IBM AWD, Palo Alto Lines: 25 In article <263@archet.UUCP> wlm@archet.UUCP (William L. Moran Jr.) writes: > >I hate to say this, but my experience is that the eafpa (enhanced >advanced ... :) is not a real big win over the regular 125 - mc68881 >I've seldom seen performance more than 20% better on pure floating >point stuff on a 135 with eafpa vs. 125 with no fpa just a mc881. I >suspect this may have something to do with the compiler as I've heard >that AIX does much better than this. Anyway, rumors I've heard say >that for AOS, the afpa may actually do worse than the mc881, although >I don't have an afpa to test this with. > > Bill Moran You're probably not seeing the best floating point performance AOS offers. The default floating-point code generated by pcc, hc, f77, and hf77 uses the run-time floating linkage (RTFL) so that objects will run on whatever floating hardware is installed. Code tailored for the run-time hardware is generated in data storage the first time a given fp operation is executed. The code is pretty good but can't match the performance of inline code. Only hf77 can generate inline code for the AFPA. Try compiling with the -f2 option, and measure that performance. Larry Breed