Path: utzoo!mnetor!uunet!husc6!bloom-beacon!gatech!purdue!i.cc.purdue.edu!j.cc.purdue.edu!pur-ee!uiucdcs!uxc.cso.uiuc.edu!urbsdc!jwilson From: jwilson@urbsdc.Urbana.Gould.COM Newsgroups: comp.sys.amiga.tech Subject: Guru 0000002f ?? Message-ID: <45300001@urbsdc> Date: 4 May 88 22:38:00 GMT Lines: 39 Nf-ID: #N:urbsdc:45300001:000:1657 Nf-From: urbsdc.Urbana.Gould.COM!jwilson May 4 17:38:00 1988 I got a very starange Guru the other day. I was running a simulation that does a ton of floating point calculcations. I cannot use the Motorolla FFP format, because I need numbers larger than 10^18. So, I was using IEEE double precision snail floating point. I use Manx C v3.6. I have an Amy 500 with 1MB ( orig 512K + Commodore 512K internal). The simulation had been running fine for about 3.5 hours ( I said it did a ton of calculations), it had produced ~400K of data in a RAM: disk. (I knew I should have run it to floppy) It was nearly finished ( I'm not sure, but it might have been at the end when it crashed), when it suddenly flashed the famous Software Failure requester. The Guru was: 0000002f.00Cxxxxx (I'm not sure of the address right now, but it was in fast ram -- where the program was running.) I have never seen a CPU Trap of 2f before, and I don't have any source that lists that number. WHAT HAPPENED??!? I dont want to waste another 4.5 hours of CPU time to have it crash again, unless I can find out what's wrong. I have successfully run this program using the FFP format with data that was smaller, but I was trying to do a larger simulation. Any help I can get from anyone would be appreciated. Jeff Wilson jwilson@urbsdc !uiucuxc!urbsdc!jwilson ******************************************************************************** working for, but in no way representing: Gould CSD-Software Division, Urbana, IL and occasionally The University of Illinois (although I pay them instead of vise versa.) ********************************************************************************