Author |
Message |
robfinch
Joined: Sat Feb 02, 2013 9:40 am Posts: 2095 Location: Canada
|
Nice proof of concept. Seems pretty decent for a late 70's computer. Quote: I had real nice post about how I planned to intoduce this computer in 1978, but the page timed out and i lost all the work. Do you have a time machine or is this something that was designed in the '70s.
_________________Robert Finch http://www.finitron.ca
|
Sun Aug 21, 2022 4:19 am |
|
|
oldben
Joined: Mon Oct 07, 2019 2:41 am Posts: 596
|
robfinch wrote: Nice proof of concept. Seems pretty decent for a late 70's computer. Quote: I had real nice post about how I planned to intoduce this computer in 1978, but the page timed out and i lost all the work. Do you have a time machine or is this something that was designed in the '70s. I planned to build some sort of computer, in the late 1970's but never had the money or access to parts or the time. Now I got the time and the internet for information. Now and then I could get to a city, with a RADIO SHACK and buy the $5 grab bag https://radioshackcatalogs.com/flipbook ... talog.htmlIn the early 1980's I got a Z80 /S100 kit, but the homade power supply toasted it, on a bad day. I did get to use a PDP-8 once . but never could buy a used PDP, so it was too collect parts and build my own computer using 2901's. With 2901's a 20 bit computer seemed like possible project, 5 slices wide rather than 4 for a 16 bit computer. After getting internet faster than 9600 baud, Home brew designs like MAGIC got me interested in them again. I was delayed about 10 years building a nice audio system, then I got Altera FPGA development card and have been prototyping with ttl logic, on them on the last few years. http://www.homebrewcpu.com/For audio https://www.enjoythemusic.com/magazine/sound_practices/ was nice place to start. With covid, parts have dried up, so I am using 16v8's replace things like 22v10's that replaced 82s100's from the late 70's. Ben.
|
Sun Aug 21, 2022 5:36 am |
|
|
BigEd
Joined: Wed Jan 09, 2013 6:54 pm Posts: 1784
|
(It's a shame to lose a drafted post, but welcome back to the days of save early and save often...)
|
Sun Aug 21, 2022 4:24 pm |
|
|
oldben
Joined: Mon Oct 07, 2019 2:41 am Posts: 596
|
BigEd wrote: (It's a shame to lose a drafted post, but welcome back to the days of save early and save often...) At 300 baud it will take a while, time for a new thread.
|
Sun Aug 21, 2022 9:08 pm |
|
|
oldben
Joined: Mon Oct 07, 2019 2:41 am Posts: 596
|
Ball park schematics and bus are done, for the version with decimal operations. Needs the use of half power PAL's that came out around 1983. Excess 3 decimal operations and Dubble Dabble adjustments, are new from 1976. 1 uS memory cycle ~1976, .65 uS memory cycle 1983. Ben.
|
Sat Sep 10, 2022 5:41 am |
|
|
oldben
Joined: Mon Oct 07, 2019 2:41 am Posts: 596
|
Adapted some older decimal floating point code, but needed a 2 more registers for the mult/divide inner loop. so I added a exchange instruction like on the z80. It works but made the design messy. Over the weekend I removed the +- 8 bit constant, freeing up a bit, so now I have up to 16 registers rather than 8 to work with. No need to have the exchange operation now. Only 12 registers fit, on the fpga how ever. I am happy with this design, and ready to clean up the software, this fall. Ben.
|
Mon Sep 19, 2022 4:57 am |
|
|
oldben
Joined: Mon Oct 07, 2019 2:41 am Posts: 596
|
Final change, BCPL freindly as I switched from little endian to big endian. Ben.
|
Wed Sep 21, 2022 2:31 pm |
|
|
drogon
Joined: Sun Oct 14, 2018 5:05 pm Posts: 62
|
oldben wrote: Final change, BCPL freindly as I switched from little endian to big endian. Ben. Are you sure? Although really, it makes no difference - the existing compiler can output code for little or big endian, but little endian is the default. (And having written a cintcode interpreter/vm in a little endian system, it all seemed to "just work" in that department) -Gordon
|
Wed Sep 21, 2022 8:30 pm |
|
|
oldben
Joined: Mon Oct 07, 2019 2:41 am Posts: 596
|
drogon wrote: oldben wrote: Final change, BCPL freindly as I switched from little endian to big endian. Ben. Are you sure? Although really, it makes no difference - the existing compiler can output code for little or big endian, but little endian is the default. (And having written a cintcode interpreter/vm in a little endian system, it all seemed to "just work" in that department) -Gordon Is this a 16 bit or 32 bit BCPL?
|
Wed Sep 21, 2022 10:57 pm |
|
|
drogon
Joined: Sun Oct 14, 2018 5:05 pm Posts: 62
|
oldben wrote: drogon wrote: oldben wrote: Final change, BCPL freindly as I switched from little endian to big endian. Ben. Are you sure? Although really, it makes no difference - the existing compiler can output code for little or big endian, but little endian is the default. (And having written a cintcode interpreter/vm in a little endian system, it all seemed to "just work" in that department) -Gordon Is this a 16 bit or 32 bit BCPL? My implementation is a 32-bit one - running on a hybrid 8/ bit CPU with an 8-bit memory bus (ie. the 65c816). The last BCPL system I used was a 16-bit one running on an 8 bit cpu (6502). (and I'm referring to Martin Richards Compiler - I suspect there are others out there but I've no idea how they handle endian) -Gordon
|
Thu Sep 22, 2022 7:35 am |
|
|
oldben
Joined: Mon Oct 07, 2019 2:41 am Posts: 596
|
drogon wrote:
My implementation is a 32-bit one - running on a hybrid 8/ bit CPU with an 8-bit memory bus (ie. the 65c816). The last BCPL system I used was a 16-bit one running on an 8 bit cpu (6502).
(and I'm referring to Martin Richards Compiler - I suspect there are others out there but I've no idea how they handle endian)
-Gordon
I will be using "Robert Nordier"'s version of BCPL. (icint.c) and Pelles C compiler under windows XYZ to get BCPL working under windows. I will later change the case statements to if statements and then translate that for my machine. Ben.
|
Thu Sep 22, 2022 9:29 am |
|
|
oldben
Joined: Mon Oct 07, 2019 2:41 am Posts: 596
|
Made a few minor changes in the hardware, and now I get routing bugs. My front panel display program is now shifting left when it is programmed to shift right. I am quickely running out midnight oil here, Ben.
|
Tue Sep 27, 2022 7:15 am |
|
|
robfinch
Joined: Sat Feb 02, 2013 9:40 am Posts: 2095 Location: Canada
|
Sounds like a direction bit is flipped.
Could be a new "feature".
_________________Robert Finch http://www.finitron.ca
|
Thu Sep 29, 2022 8:09 am |
|
|
oldben
Joined: Mon Oct 07, 2019 2:41 am Posts: 596
|
I could not solve the problem, so I had to roll back to a earlier version that used less block ram. I am leaving the instruction set as is. and plan to work on external I/O. I planing a 16 bit hexadecimal front panel rather than a binary one, and a new setup for serial and SD cards. Now if I could only find 14 segment NIXIE TUBES rather than the common 7 segment led's. Ben.
|
Thu Sep 29, 2022 12:08 pm |
|
|
Ken KD5ZXG
Joined: Sat Sep 03, 2022 3:04 am Posts: 51
|
ин-14 nixie tubes? But these are not segmented. Or 14 segmented vacuum fluorescent display, somehow in nixie red? Maybe from some unrepairable chess computer, last place I saw anything like that... Go big with orange/red EL ropelights.
|
Thu Sep 29, 2022 9:11 pm |
|
|