Hi Martin, Still have an Alpha in storage that still has the software that was used to control a hot strip mill (for a steel company). When I was still consulting my role was writing software for communications with PLCs and integrating it with software (fortran) that another wrote (an engineer that understood the physics involved) that controlled the steel rollers to keep the steel sheets within a certain tolerance. This was actually a ML (Machine Learning) application since the software was able to learn from the data from the PLCs and make the necessary adjustments.
Again something to read when you are having trouble sleeping. John S On Tue, Jan 13, 2026 at 2:43 PM Martin Kobetic <[email protected]> wrote: > Hi John, interesting stuff. We are getting a bit off-topic, hopefully > others don't mind. I joined the party in the 90s, I remember HP pa-risc, > Sun Sparc, SGI Indigo workstations. Intel and PCs were just establishing > their foothold. I haven't really worked with the Alphas. It sounds like you > have a lot of material for some interesting blog posts if you were so > inclined. > > Best regards, > Martin > > On Tue, Jan 13, 2026 at 12:54 PM John Sarabacha <[email protected]> > wrote: > > > Hi Again Martin, > > These emails are for when you have trouble sleeping (will put you > asleep). > > Bill Gates hired Dave Cutler away from DEC (who worked on vms 32 bit > > operating systems) to bring forth (no pun intended) Windows NT (their 32 > > bit version of windows) and it carried on from there to this day. So > > software engineering is a surviving skill through time. Hope some of the > > younger readers of these emails understand this, you work with hardware > but > > don't get constrained by it. > > > > Regards, > > John S > > > > > > On Tue, Jan 13, 2026 at 12:07 PM John Sarabacha <[email protected]> > > wrote: > > > > > Hi Again Martin, > > > How things have changed, when I worked at DEC (PDP11s, VAX, Alpha ...) > > > many years ago where are they (DEC) now? > > > Part of HP now, did you know that DEC was offered to own unix (turned > it > > > down). They were also the 1st to use expert systems to configure PDP, > > Vax, > > > microVax, ... hardware configurations. HP was the 1st to bring out RISC > > > processors and DEC responded with their 64 bit RISC Alphas. Now the > > > chinese are dominating the RISC market using royalty-free RISCV core > > chips, > > > ARM (risc) is still ahead for the time being in high end processors. > The > > > Cortex-M7 platform that I referenced before is approx 35$us and it > easily > > > outperforms the dated DEC Alpha which at one time ran the internet > search > > > engine of choice. Just ranting on in my so-called retirement years. > > > > > > Best Regards, > > > John S > > > > > > > > > > > > On Tue, Jan 13, 2026 at 10:07 AM John Sarabacha <[email protected]> > > > wrote: > > > > > >> Hi Again Martin, > > >> A lot of useful information from that link, thanks again. There is a > > >> temptation to use C for these core (foundation words - mcu assembly > > words) > > >> since compilers these days can generate very efficient code. The issue > > is > > >> amForth tries to be compatible with the 2012 forth standard whereas > > other C > > >> based forths like ueForth don't. It still makes more sense just to > make > > the > > >> adjustments to amForth assembly code keeping in mind the C forth code > > >> (looked at) compiled with -S (for the assembler code) may deviate from > > this > > >> standard. > > >> > > >> Best Regards, > > >> John S > > >> > > >> > > >> > > >> > > >> > > >> > > >> On Tue, Jan 13, 2026 at 9:15 AM John Sarabacha <[email protected]> > > >> wrote: > > >> > > >>> Thanks Martin, > > >>> That information saves me a lot of time, what I was also thinking of > > >>> doing is letting the compiler tell me > > >>> how to be compatible, this is where ueForth (being in C) can help, > for > > >>> each C word compile with -S and > > >>> see what the assembly listing looks like and compare this to what the > > >>> amForth word has and make the adjustments. > > >>> > > >>> Thanks again, > > >>> John S > > >>> > > >>> On Tue, Jan 13, 2026 at 8:52 AM Martin Kobetic <[email protected]> > > >>> wrote: > > >>> > > >>>> Hi John, > > >>>> That is a sweet looking board. I need to resist the urge to get one > > too. > > >>>> I've got plenty that I didn't even use yet :). > > >>>> > > >>>> I'm not that well versed in ARM yet, but the errors sound like the > > >>>> instruction mode > > >>>> is being switched around for some reason. Maybe we need to ensure > > >>>> somehow > > >>>> that > > >>>> it sticks with Thumb? There may also be some issue with > interworking, > > I > > >>>> haven't quite > > >>>> wrapped my head around that either. A lot to do still. > > >>>> > > >>>> Some interesting references > > >>>> > > >>>> > > > https://stackoverflow.com/questions/28669905/what-is-the-difference-between-the-arm-thumb-and-thumb-2-instruction-encodings > > >>>> > > >>>> > > > https://developer.arm.com/documentation/dui0203/j/interworking-arm-and-thumb/about-interworking > > >>>> > > >>>> Best regards, > > >>>> Martin > > >>>> > > >>>> On Tue, Jan 13, 2026 at 12:40 AM John Sarabacha < > [email protected] > > > > > >>>> wrote: > > >>>> > > >>>> > Hi Martin, > > >>>> > The Cortex-M7 doesn't like the assembler instructions for > arm/words, > > >>>> some > > >>>> > examples: > > >>>> > ./words/dup.s:4: Error: Thumb does not support this addressing > mode > > >>>> -- `str > > >>>> > tos,[psp,#-4]!' > > >>>> > ./words/nip.s:4: Error: instruction not supported in Thumb16 mode > -- > > >>>> `adds > > >>>> > psp,#4' > > >>>> > ./words/abs.s:4: Error: Thumb does not support conditional > execution > > >>>> > ./words/abs.s:5: Error: incorrect condition in IT block -- `b > > DO_NEXT > > >>>> > > > >>>> > Tried different options for the GCC compiler/assembler without > > >>>> success. > > >>>> > > > >>>> > FYI > > >>>> > John S > > >>>> > > > >>>> > On Mon, Jan 12, 2026 at 1:13 PM John Sarabacha < > > [email protected]> > > >>>> > wrote: > > >>>> > > > >>>> > > Hi Martin, > > >>>> > > Some information on my Cortex-M7 based target platform, Arduino > > >>>> > > development environment is being is used > > >>>> > > https://www.pjrc.com/store/teensy41.html > > >>>> > > > > >>>> > > Regards, > > >>>> > > John S > > >>>> > > > > >>>> > > On Mon, Jan 12, 2026 at 12:04 PM John Sarabacha < > > >>>> [email protected]> > > >>>> > > wrote: > > >>>> > > > > >>>> > >> Hi Martin, > > >>>> > >> I will be trying to host amForth on an arm-m7 (600Mhz) platform > > >>>> which > > >>>> > was > > >>>> > >> being evaluated a few years ago. I will be using the same > process > > >>>> as I > > >>>> > used > > >>>> > >> for CH32X033. Any learning I will pass on, M7 is significantly > > >>>> different > > >>>> > >> than M4 so by sticking to thumb type instructions there should > be > > >>>> some > > >>>> > >> common issues. > > >>>> > >> > > >>>> > >> Regards, > > >>>> > >> John S > > >>>> > >> > > >>>> > > > >>>> > > > >>>> > > >>>> _______________________________________________ > > >>>> Amforth-devel mailing list for http://amforth.sf.net/ > > >>>> [email protected] > > >>>> https://lists.sourceforge.net/lists/listinfo/amforth-devel > > >>>> > > >>> > > > > _______________________________________________ > > Amforth-devel mailing list for http://amforth.sf.net/ > > [email protected] > > https://lists.sourceforge.net/lists/listinfo/amforth-devel > > > > _______________________________________________ > Amforth-devel mailing list for http://amforth.sf.net/ > [email protected] > https://lists.sourceforge.net/lists/listinfo/amforth-devel > _______________________________________________ Amforth-devel mailing list for http://amforth.sf.net/ [email protected] https://lists.sourceforge.net/lists/listinfo/amforth-devel
