Path: utzoo!utgpu!watmath!att!tut.cis.ohio-state.edu!brutus.cs.uiuc.edu!wuarchive!wugate!uunet!mcvax!sunic!kth!draken!bmc1!kuling!irf From: irf@kuling.UUCP (Bo Thide') Newsgroups: comp.sys.hp Subject: Re: Floating point accelerator 98248a Message-ID: <1137@kuling.UUCP> Date: 16 Aug 89 21:57:31 GMT References: <112@engr.wisc.edu> <5570266@hpfcdc.HP.COM> Reply-To: irf@kuling.UUCP (Bo Thide') Organization: Dept. of Computer Systems, Uppsala University, Sweden Lines: 30 In article <5570266@hpfcdc.HP.COM> rjn@hpfcdc.HP.COM (Bob Niland) writes: >re: wittmann@engr.wisc.edu (art wittmann) writes.... > >> I just got a 98248a FPA for my 9000/360. >> ... >> ...saw about a 10% performance increase. > >The recommended FPA for the 360/370 is the 98248B. The 98248A was >optimized for the 330 and 350. Offhand, I'm not sure if the 10% result >is typical for your configuration. > I had a 9828A for my 350 and saw a 100-250% performance increase in a very floating-point oriented test. When I upgraded to the 370 I also changed to the 98248B (Dragon II) and saw a somewhat lower performance increase. This is probably due to the fact that the 33MHz 68882 in the 370 is much faster than the 20 MHz 68881 that I had in the 350 so the difference between the coprocessor and the FPA is smaller in the 370. Of course, the 370 + Dragon II is a faster floating point machine than the old 350 + Dragon I but not as much as I had hoped. I understand the 300 series FPAs are manufactured by Intel (?). Why didn't HP choose to replace Dragon I with a similar PFA as in the 9000/835 which is *much* faster? ^ Bo Thide'-------------------------------------------------------------- | | Swedish Institute of Space Physics, S-755 91 Uppsala, Sweden |I| [In Swedish: Institutet f|r RymdFysik, Uppsalaavdelningen (IRFU)] |R| Phone: (+46) 18-403000. Telex: 76036 (IRFUPP S). Fax: (+46) 18-403100 /|F|\ INTERNET: bt@irfu.se UUCP: ...!uunet!sunic!irfu!bt ~~U~~ -----------------------------------------------------------------sm5dfw