Re: [Simh] Interdata OS/32: hello-world in CAL32
Thanks very much for the additional info. Your post was very timely since I read in the notes that come with the PL/M cross compiler that is being discussed that it was a cross-compiler hosted on MTS and VM/CMS. I don't think I ever came across a cross compiler in the old days. It is interesting to see that people used these odd combinations. I wonder if we should start trying to archive and document cross compilers specifically. On Tue, 23 Feb 2016 01:01:28 -0500 johns...@gregjohnson.org wrote: [very nice story snipped] ___ Simh mailing list Simh@trailing-edge.com http://mailman.trailing-edge.com/mailman/listinfo/simh
Re: [Simh] Interdata OS/32: hello-world in CAL32
On Feb 23, 2016, an anonymous user (li...@openmailbox.org) wrote: > Thanks very much for the additional info. Your post was very timely since I > read in the notes that come with the PL/M cross compiler > that is being discussed that it was a cross-compiler hosted on MTS and VM/CMS. > I don't think I ever came across a cross compiler in the old days. It is > interesting to see that people used these odd combinations. > I wonder if we should start trying to archive and document cross compilers > specifically. > On Tue, 23 Feb 2016 01:01:28 -0500 > johns...@gregjohnson.org wrote: > [very nice story snipped] Cross compilers were great when your target system didn't have enough memory, storage, and possibly no actual operating system to support a locally-hosted compiler. For a language like PL/M-86 which had no language specific I/O constructs beyond reading to and writing from I/O ports on the target CPU, it was perfect for writing applications for an OS-less embedded systems, but it needed a host. Hosting on a VAX under VMS might be viewed as a little extreme when the target had less than 64K of EPROM and a few K of RAM, but it worked. The only complaint I had is that Intel limited the symbol table sizes somewhere in the linker (IIRC !), and you could exceed the maximum number of external symbols, thus requiring some splitting of modules and multi-step linking to resolve this dilemma. I suspect the origins of this limitation was the MS-DOS environment with 640K of memory (Intel's tools didn't use a DOS extender), and that whoever ported the toolset to VAX/VMS never increased this limitation, even though VAX/VMS could support a lot more than virtual memory than DOS. Assemblers were a little bit easier to host on a target - smaller code size in your editor, no optimization required, etc., though plenty of cross assemblers certainly existed for a wide range of targets, especially those that were embedded like Intel 8048 & 8051, Motorola 680x, etc. A lot of the "classic" embedded cross compiler/cross assembler companies are no longer in existence, either gobbled up by larger companies, or simply going the way of the Dodo bird. Names like Franklin, 2500AD, Avocet System, Adtek (from Japan) and Hi-Tech (from Queensland Australia), were pretty common players, but these days their products are no longer available, or if you're lucky they are end-of-life and provided as-is with no support, even if you buy them. Intel created their own systems, like the Intel's iPDS-100 running ISIS-II, or Motorola/Freescale (now NXP) got 3rd party vendors like P&E Micro to do some of their tools and development boards. It was a wild time !!! PS: I'd love to see an iPDS-100 emulation in SIMH one day !!! ___ Simh mailing list Simh@trailing-edge.com http://mailman.trailing-edge.com/mailman/listinfo/simh
Re: [Simh] Interdata OS/32: hello-world in CAL32
On Tue, 23 Feb 2016 14:15:11 + "Armistead, Jason BIS" wrote: > Hosting on a VAX under VMS might be viewed as a little extreme when the > target had less than 64K of EPROM and a few K of RAM, but it worked. That's what I meant. I know what a cross compiler is. I never saw a mainframe-hosted cross-compiler back in the day and here within 2 days I found out about two mainframe hosted cross compilers. And then your note about extreme re: VAX/VMS. Imagine! ___ Simh mailing list Simh@trailing-edge.com http://mailman.trailing-edge.com/mailman/listinfo/simh
Re: [Simh] Cross Compilers (and memories thereof)
Let's move this to a new thread subject of its own ! On Tuesday, 23 February 2016 12:04 PM, Davis Johnson wrote (under old subject Re: [Simh] Interdata OS/32: hello-world in CAL32) : > One that I remember was TI had a 9900 cross assembler written in FORTRAN (all > caps in those days). It was free to educational institutions. > I talked a prof. into requesting it, but the available FORTRAN compiler > didn't like it. Another now-defunct company, Microtec Research Incorporated (purchased by Mentor Graphics in late 1995), definitely had a TI9900 cross assembler written in FORTRAN from 1983. We had the source code under license and compiled it under VAX/VMS (complete with CLI switches), and with a few custom tweaks, it was largely compatible with SDSMAC that ran on TI's 990 computer systems (now simulated via Dave Pitts' SIM990). There was also a linker/loader that produced Tektronix HEX output (similar to TI's SDSLNK) as the final executable. I was able to modify the code enough to get it to compile under OpenWatcom's Fortran 77 on Windows XP. ___ Simh mailing list Simh@trailing-edge.com http://mailman.trailing-edge.com/mailman/listinfo/simh
Re: [Simh] Cross Compilers (and memories thereof)
I note someone said MTS hosted several cross assemblers and compilers. Readers may be interested to note that a sanitised copy of MTS ready to run on the Hercules S390/XA/390 emulator is available for download. http://archive.michigan-terminal-system.org/mts-d60A http://www.hercules-390.eu/ Dave Wade G4UGM > -Original Message- > From: Simh [mailto:simh-boun...@trailing-edge.com] On Behalf Of > Armistead, Jason BIS > Sent: 23 February 2016 17:32 > To: simh@trailing-edge.com > Subject: Re: [Simh] Cross Compilers (and memories thereof) > > Let's move this to a new thread subject of its own ! > > On Tuesday, 23 February 2016 12:04 PM, Davis Johnson > wrote (under old subject Re: [Simh] Interdata OS/32: hello-world in CAL32) : > > > One that I remember was TI had a 9900 cross assembler written in > FORTRAN (all caps in those days). It was free to educational institutions. > > I talked a prof. into requesting it, but the available FORTRAN compiler > > didn't > like it. > > Another now-defunct company, Microtec Research Incorporated (purchased > by Mentor Graphics in late 1995), definitely had a TI9900 cross assembler > written in FORTRAN from 1983. We had the source code under license and > compiled it under VAX/VMS (complete with CLI switches), and with a few > custom tweaks, it was largely compatible with SDSMAC that ran on TI's 990 > computer systems (now simulated via Dave Pitts' SIM990). There was also a > linker/loader that produced Tektronix HEX output (similar to TI's SDSLNK) as > the final executable. I was able to modify the code enough to get it to > compile under OpenWatcom's Fortran 77 on Windows XP. > > > > > > ___ > Simh mailing list > Simh@trailing-edge.com > http://mailman.trailing-edge.com/mailman/listinfo/simh ___ Simh mailing list Simh@trailing-edge.com http://mailman.trailing-edge.com/mailman/listinfo/simh
Re: [Simh] Cross Compilers (and memories thereof)
Indeed, the PL/M compiler comes with notes saying it was hosted on MTS. This thread could be the start of something very interesting! Thanks guys! On Tue, 23 Feb 2016 17:39:58 - "Dave Wade" wrote: > I note someone said MTS hosted several cross assemblers and compilers. > Readers may be interested to note that a sanitised copy of MTS ready to > run on the Hercules S390/XA/390 emulator is available for download. > > http://archive.michigan-terminal-system.org/mts-d60A > > http://www.hercules-390.eu/ > > Dave Wade > G4UGM > > > -Original Message- > > From: Simh [mailto:simh-boun...@trailing-edge.com] On Behalf Of > > Armistead, Jason BIS > > Sent: 23 February 2016 17:32 > > To: simh@trailing-edge.com > > Subject: Re: [Simh] Cross Compilers (and memories thereof) > > > > Let's move this to a new thread subject of its own ! Indeed! :-) > > > > On Tuesday, 23 February 2016 12:04 PM, Davis Johnson > > wrote (under old subject Re: [Simh] Interdata OS/32: hello-world in > > CAL32) : > > > > > One that I remember was TI had a 9900 cross assembler written in > > FORTRAN (all caps in those days). It was free to educational > > institutions. > > > I talked a prof. into requesting it, but the available FORTRAN > > > compiler didn't > > like it. > > > > Another now-defunct company, Microtec Research Incorporated (purchased > > by Mentor Graphics in late 1995), definitely had a TI9900 cross > > assembler written in FORTRAN from 1983. We had the source code under > > license and compiled it under VAX/VMS (complete with CLI switches), and > > with a few custom tweaks, it was largely compatible with SDSMAC that > > ran on TI's 990 computer systems (now simulated via Dave Pitts' > > SIM990). There was also a linker/loader that produced Tektronix HEX > > output (similar to TI's SDSLNK) as the final executable. I was able to > > modify the code enough to get it to compile under OpenWatcom's Fortran > > 77 on Windows XP. ___ Simh mailing list Simh@trailing-edge.com http://mailman.trailing-edge.com/mailman/listinfo/simh
Re: [Simh] Cross Compilers (and memories thereof)
On 23-Feb-16 12:32, Armistead, Jason BIS wrote: > Let's move this to a new thread subject of its own ! > > On Tuesday, 23 February 2016 12:04 PM, Davis Johnson > wrote (under old subject Re: [Simh] Interdata OS/32: hello-world in CAL32) : > >> One that I remember was TI had a 9900 cross assembler written in FORTRAN >> (all caps in those days). It was free to educational institutions. >> I talked a prof. into requesting it, but the available FORTRAN compiler >> didn't like it. > Another now-defunct company, Microtec Research Incorporated (purchased by > Mentor Graphics in late 1995), definitely had a TI9900 cross assembler > written in FORTRAN from 1983. We had the source code under license and > compiled it under VAX/VMS (complete with CLI switches), and with a few custom > tweaks, it was largely compatible with SDSMAC that ran on TI's 990 computer > systems (now simulated via Dave Pitts' SIM990). There was also a > linker/loader that produced Tektronix HEX output (similar to TI's SDSLNK) as > the final executable. I was able to Tektronix had a VMS-hosted cross-compiler & IDE for the 68000 that was used by a number of DEC communications projects. And one of mine. I didn't have sources, but I did have an interesting time with it. Replacing some RTL routines to allow multitasking. And then there was the problem of where it put string literals. I ended up having the C compiler output assembly, then inserting a processing step to move the literals to ROM, eliminate duplicates, consolidate substrings -- and then let the assembler and linker do their things. One could get almost position-independent code out of it, so with some more out-of-band fixups, I could get tasks to load from the (flash) filesystem... I ended up writing a custom OS that had a rather similar API to VMS - QIOs, event flags, ASTs - all in an unmapped 68000. I shared code written in C between VMS device drivers/user level and the embedded system. Not many #ifdefs. I did use 32 bits in some structures that VMS skimped on. But SDL was really handy for keeping the definitions consistent between two 32-bit and two 8-bit architectures Microtec Research bought the cross-compiler suite from Tek sometime in the 90s. Tek was in one of those "consolidating to core business" phases. Of course, their core business was the world's best CRTs. (For oscilloscopes.) smime.p7s Description: S/MIME Cryptographic Signature ___ Simh mailing list Simh@trailing-edge.com http://mailman.trailing-edge.com/mailman/listinfo/simh