Okay, here's the screenshot of the #PERQ emulator holding up today's newspaper. For fans of #CommonLisp or #retrocomputing in general, I've slapped together an archive here: https://github.com/skeezicsb/PERQmedia/tree/master/Accent_S6/Lisp_M3 with a pre-configured hard disk image and config files. The T2 image works with the current release (v0.7.5) but to run the T4/4MB configuration requires a build from source or an updated release (v0.7.8) coming Real Soon Now. [Edited: oh yeah, Hemlock runs too.] Share and enjoy!
#perq
0 posts0 participants0 posts today
#PERQ Emulator progress: 24-bit/4MB "T4" configuration now booting Accent S7. Oooooooooooh can a Spice Lisp with enough space to do some stuff be far behind? :-)
Time to cast a wider net, but I'm not a fan of the "throw 58 hashtags at it to see what sticks" shotgun approach, so I'll just ask the handful of #PERQ nuts still around to look to their left, look to their right, and point out the person who might have the PNX sources or documentation I seek. (If it's neither of them, maybe it's YOU! Go get that dusty old box of 8" floppies from the back of that shelf and bust out the Catweasel.)
Because surely someone, somewhere, remembers SERC, the 1980s, Chilton, RAL, ICL Dalkieth, PNX, etc. -- and squirreled away some docs, or diskettes, or even a whole PERQ with *some* information about the "C-Codes" instruction set and virtual C machine architecture of PNX. NMOC? Bletchley Park? Beuller? Anyone? There can't possibly still be concerns about throwing a copy of the sources over the transom 40 years on for research and preservation.
In the meantime, reverse engineering the kernel's microcode debugger is arduous and painstaking. Here's where we're at after a floppy or hard disk boot: both halt in a similar manner, somewhere at or before attempting to tekload the updated Z80 firmware. I'll get there, eventually, through sheer doggedness. But PNX kernel hackers of yore, lurking in #vintagecomputing, long free of ICL's closed-source bloody-mindedness, chime in anytime. :-)
#Smalltalk -80 on #PERQ --- Mario Wolczko's implementation running with maybe some gremlins still to sort out. attn @skeezicsb as usual 
Replied in thread
#PERQ documentation for cc may not talk about -S, and the assembler may be classified, but cc -S still works just fine. It should not be too hard to figure out more details of the architecture by looking at compiled C code. attn: @skeezicsb
Continued thread
#PERQ PNX seems a bit cagey about its own instruction set! cc(1) lists no -S flag for generating assembly output, but it does mention an assembler. OK, so let's man as:
Here's the manual for the ICL PNX C compiler. PNX, a System III UNIX for #PERQ workstations, uses its own instruction set, which it programs into the PERQ's CPU microcode. But there is a mysterious absence...
Continued thread
It’s video!! But something is obviously wrong. The patterns suggested a flaw in the dedicated “RasterOp” hardware the #PERQ uses to move blocks of memory, both for video and for internal OS stuff. And at 231, POS is using RasterOp!
Continued thread
My #PERQ monitor needs repair for a part that emits 20000 volts of DC power. I’m not ready to work on that yet! So I went on a new side quest: convert PERQ video to VGA. This adapter took some work, but it’s doing it. There’s a 70 MHz signal passing through that mess! Schematic eventually…
Continued thread
The original #PERQ OS is “POS”, but in the UK, a Unix called “PNX” (yes, odd names) was preferred. When PNX boots, it only counts up to 255. Imagine my astonishment when it got that far on the first try: my PERQ had booted an operating system! And my oscilloscope found something new: a video signal…
Continued thread
Continued thread
And talking of side quests: I went back to Somerset to collect these hard drives, which hold copies of the very rare #PERQ Flex operating system. Recovering data from these drives is a challenge for another month. It’s an honour to be trusted with them.
Continued thread
#PERQ 2 machines have this diagnostic counter that counts up during boot. If it stops counting before some OS-specific number in the hundreds, there was a problem. When #RetroChallenge 2024 began, the counter was saying 010. Basic tests pass, but disks don’t work. No boot.
Continued thread
All those chips make #PERQ infamous to restorers: so much to break. And there are two more boards: memory/video and I/O! To top it off, my 2nd-gen PERQ lived many years in a damp Somerset cellar, and looks it. TBH I expected things to be much worse: lots of this PERQ was working. But plenty wasn’t.
Continued thread
#PERQ had Alto-like perks. Bitmap display. Mouse (or a tablet at least). Networking. Programmable microcode: program the CPU for YOUR favourite instruction set. The designers had Pascal code in mind. No late-70’s microprocessor was enough: Three Rivers made its own CPU from LOTS of discrete logic.
Continued thread
#PERQ was made by a startup, Three Rivers Computer in Pittsburgh. They soon partnered with the UK firm ICL, who wanted to make scientific workstations for British labs and universities. It’s a whole saga.
Continued thread
It’s just turned November here — that wraps up #RetroChallenge 2024! I dedicated my month to reviving an early 1980s #PERQ 2T2 graphical workstation. Let me show you how far I got… 
Replied in thread
@andreas_heitmann @smittytone I sincerely hope you’re not taking me for an #Apple fanboy. I am not. I just happen to be using #UNIX-like OSs ever since 1984 (the first one being #PNX on an #ICL #Perq — Learn all about this one on Wikipedia. This includes #NetBSD and assorted #Linux based sytems. I just believe one has to be flexible enough not to bet fixated on a particular feature perhaps better handled by another system. Apple isn’t doing badly at all.
Say you want to turn the #PERQ HSYNC and PERQ VSYNC signals shown into the VGA HSYNC and VGA VSYNC signals. (TTL levels.) How do you do it? 74LS123 has been proposed; could try a small PIC but I haven't programmed a PIC before. #electronics #RetroChallenge sidequest.