jgilbert@wsmr08.ARPA (10/26/84)
From: John Gilbert CD <jgilbert@wsmr08.ARPA> I wrote a program in Turbo Pascal that runs fine on a Northstar, a KayPro II and an Osborne I. I recently tried to run it on an Osborne Executive and it gives all kind of extraneous results when I am doing recursive computation. I did not change to upper memory bound and am under the impression the TPA of the Executive is much larger than that of the Osborne I. I did learn that the Executive uses something called CPM+ and subsequently learned that it has 128K of RAM and bank swaps. However, from what I could read that would not appear to cause the phenomena I encountered. Can anyone provide me any insight into potential incompatibilities between the Osborne Executive and the other CPM2.2 computers I mentioned or a suggestion about what might be causing my problem? John
W8SDZ@Simtel20.ARPA (10/27/84)
From: Keith Petersen <W8SDZ@Simtel20.ARPA> Hello, John. I would run a memory test first, to make sure the new Osborne is working correctly. We just got a new one: Filename Type Bytes CRC Directory MICRO:<CPM.MEMTEST> MEMTST10.ASM.1 ASCII 14683 81A0H MEMTST10.COM.1 COM 1024 D129H MEMTST10.HEX.1 ASCII 2516 D80FH There are others in that directory, including WORM21, which is a program that tests memory-access timing problems on the Z80. The only other thing that comes to mind is that perhaps the Osborne CBIOS is clobbering some of the Z80 alternate registers and maybe Turbo Pascal is using them. Used to be that program writers could assume that the CP/M system would not touch anything but the 8080 registers and so you automatically save them when calling BDOS. Nowadays it seems that many OEMs are using the alternate and/or index registers of the Z80, thus destroying their contents as far as your program is concerned. One popular Public-Domain spelling checker had to be re-written because of this. --Keith