Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!usc!snorkelwacker!bloom-beacon!eru!luth!sunic!mcsun!unido!mpirbn!p554mve From: p554mve@mpirbn.UUCP (Michael van Elst) Newsgroups: comp.sys.amiga.hardware Subject: Re: 68881 in 68000 system Message-ID: <947@mpirbn.UUCP> Date: 3 May 90 10:02:50 GMT References: <3676@minyos.xx.rmit.oz> Reply-To: p554mve@mpirbn.UUCP (Michael van Elst) Organization: Max-Planck-Institut fuer Radioastronomie, Bonn Lines: 19 In article al158305@mtecv2.mty.itesm.mx (Gustavo Cordova) writes: [about the 68881 working with an 68000] >yes, I think that you need that math.resource, and that the chip can't be >used when multitasking programs that use the chip, I think that because >the OS doesn't/can't save the FPU registers...but, then again, I can be >wrong :) You need a minimal math.resource that merely says where the 68881 is mapped in memory. All processes that want to access the chip may do this via the mathieeedoubbas.library or mathieeedoubtrans.library. Each process has to open the library itself so that the library can install a task-specific routine that saves and reloads the 68881 registers. -- Michael van Elst UUCP: universe!local-cluster!milky-way!sol!earth!uunet!unido!mpirbn!p554mve Internet: p554mve@mpirbn.mpifr-bonn.mpg.de "A potential Snark may lurk in every tree."