You realise the server runs at 8mhz right? You realise that there is no room for extra there?
Perhaps you would like to show us your work on a website that runs on an 8mhz processor?
Walter Belgers and Marcel van Kervinck's Gigatron (https://gigatron.io/) is also interesting. It's spirit is to use only 70's technology to keep the system "learnable" from the ground up, although they sometimes violate this rule for peripherals outside the main system.
In the age of GitHub pages, Netlify and who knows how many other unlimited bandwidth web hosts - how on earth do people still manage to get their sites slashdotted?
This is definitely on my "list of things I've always wanted to do but I realized man is mortal and I have to prioritize". I'm a personal computing enthusiast and it doesn't get much more personal than designing and building your own architecture.
Before I saw the German, I already had a feeling it was --- I'm not sure why, but I've seen Germans use that style of page a lot (another example: https://www-user.tu-chemnitz.de/~heha/basteln/PC/USB2LPT/ul-... ). That and the encouragement to build one yourself stood out.
This one looks very similar to a 6502, and seems to be much smaller than another semi-famous discrete homebrew computer, http://www.homebrewcpu.com/ --- that one only runs at half the clock speed, but the performance seems comparable.
I did this for a living in the 1980s--CPUs from TTL. My artistic side wants to build a small running computer from all TTL inside a plexiglas cube with all the chips glued to the walls and the interconnecting multi-colored wires exposed with a small display of something.
He even sells a kit with all parts and schematics included. It's easily one of my favorite projects of all time and has inspired me to continue learning more.
It's not close to MyCPU's level of complexity but is great to get you started.
I took a compilers course from Prof. Porter, and at the time he invited undergraduates to write a simulator for his relay computer. I remember thinking, "why in the world would I want to do that? I want to work on the next big thing!" Fast forward to now, I've never worked on the Next Big Thing, and am instead slowly, though enjoyably, building Ben Eater's 8-bit SAP-1 when I have some free time. It sort of feels a bit like building a ship in a bottle. I just wish I had taken Professor Porter up on his offer and maybe scratched an itch when I had a little more energy and free time.
I have been looking at your board on Tindie! Alas, I don't think it should be my first major soldering job (if you don't count the board I stuffed 25 odd years ago that my friend had to fix for me)
I love these homebrew CPU projects. I saw one one in person and was amazed (and impressed by the insane dedication as well as the madness of wire-wrapping 1970s-era TTL components into a CPU) - then a year later I saw two of them!! Then I saw a third.
I believe the first two I saw were MAGIC-1 and BMOW (Big Mess o'Wires). I also really liked the MOnSter 6502, a beautiful board-level reimplementation of the 6502 in surface-mount discrete components, with LEDs on the data lines and optional slow/single-step clocking.
Also relevant: the nand2tetris.org course, and Niklaus Wirth's TRM/RISC-3 (not to be confused with Berkeley RISC-I or modern RISC-V) FPGA implementation. Notably TRM/RISC-3 is incredibly simple but can run the whole Oberon system.
In the modern era, I think FPGAs are the killer platform for computer architecture experimentation. There are multiple free MIPS implementations for FPGA and of course there is great OS/compiler support for MIPS.
A nice project, there are more homebrew CPUs - physical or FPGA cores. But they are often complex. I was finally able to understand the basics of CPU workings from an awesome book http://www.buthowdoitknow.com/ and related video https://youtu.be/cNN_tTXABUA
40 comments
[ 3.2 ms ] story [ 83.6 ms ] threadI would imagine now that it's on the front page of HN, that poor MyCPU is now on fire
He has remade a system similar to the core of the great 1970's Xerox Alto / Xerox Star series :-)
I never thought someone would remake a processor this way. I love that kind of madness !
[edit: use the site backup if it fails to load: http://mycpu.thtec.org/www-mycpu-eu/index1.htm ]
https://hackaday.io/projects?tag=CPU (Some are FPGA, most are TTL) Including two risc-v implementations (one seems stalled):
https://hackaday.io/project/18491-worlds-first-32bit-homebre... (No activity for a while now) and https://www.youtube.com/playlist?list=PLEeZWGE3PwbansoxKjjMK...
One I am particularly interested in is the Gigatron:
https://hackaday.io/search?term=Gigatron&category=project%20... (Especially now that there's real IO being hacked in)
Walter gave a talk recently about it: https://www.youtube.com/watch?v=rbaAmnhZAis
[1] http://web.archive.org/web/20190516084513/http://mycpu.eu/
Nobody mentions this.
In every wire rat nest CPU picture, there is always several thousand dollars of digilent and HP gear out of the frame.
This one looks very similar to a 6502, and seems to be much smaller than another semi-famous discrete homebrew computer, http://www.homebrewcpu.com/ --- that one only runs at half the clock speed, but the performance seems comparable.
https://eater.net/8bit
He even sells a kit with all parts and schematics included. It's easily one of my favorite projects of all time and has inspired me to continue learning more.
It's not close to MyCPU's level of complexity but is great to get you started.
Here’s a real computer. http://web.cecs.pdx.edu/~harry/Relay/
http://relaysbc.sourceforge.net/
(apologies for the self promotion..)
I believe the first two I saw were MAGIC-1 and BMOW (Big Mess o'Wires). I also really liked the MOnSter 6502, a beautiful board-level reimplementation of the 6502 in surface-mount discrete components, with LEDs on the data lines and optional slow/single-step clocking.
Also relevant: the nand2tetris.org course, and Niklaus Wirth's TRM/RISC-3 (not to be confused with Berkeley RISC-I or modern RISC-V) FPGA implementation. Notably TRM/RISC-3 is incredibly simple but can run the whole Oberon system.
In the modern era, I think FPGAs are the killer platform for computer architecture experimentation. There are multiple free MIPS implementations for FPGA and of course there is great OS/compiler support for MIPS.