Home Page Forums MAME MAME Intergraph InterPro WIP

It's not that scary, there's just a small matter of programming it, is all, taking care of all ROPs as well.

For my latest trick, I now present the InterPro "blue screen", the entry point for configuring the system and formatting disks and installing CLIX etc.

The mouse pointer moves, but doesn't track vertically for some reason, and there are also some lines missing (e.g. around the buttons) because my line drawing hardware emulation isn't up to scratch yet, but this is a milestone for me.

The toughest part was working out how the "data select" mode operates when writing to video buffer memory - essentially it expects a write to video memory of 4 bits to expand to 4 8bpp pixels/bytes using the active "plane data". Of course, now that I know/understand it's all obvious

Very, very cool.

Considering how aggressively secretive I remember Intergraph being back in the 80s/90s I'm really impressed it's gotten this far.

Got the mouse tracking properly, and enough line drawing to make it usable. Time to start playing with SCSI and such again.

Was there an x86 card available for those workstations to run DOS, or is it an emulator?

In fact, there was both a hardware x86 emulation card, named "Hard PC", as well as a port of the relatively well-known SoftPC software emulator. Unfortunately, I have next to no information on the former; a quick search throws up the link below, which is interesting, but doesn't reveal anything too informative about the hardware.

I'm not at all sure if the PC-DOS button on the blue screen was the only way to boot this Hard PC hardware, or whether it could be accessed concurrently via CLIX. I do know that there was an InterPro model 29xx system designed and manufactured which was based on the 2xxx series case/monitor etc., but used a PC (Pentium?) motherboard, but this system was not able to run CLIX or CLIPPER software. It was basically just a PC in an InterPro box, designed to run Windows NT, supposed to allow existing customers to extend the life of their InterPro monitors and maybe some other peripherals.

The Hard PC board would be an excellent longer term emulation target, but I suspect there's even less information available than the slim pickings I have for the InterPro, so progress may be very slow.

Just found a few more bits and pieces of information - it appears the Hard PC had a 486 CPU, some kind of VGA hardware/emulation, and supported 2, 4, 8 or 16 megabytes of its own SIMM memory. It appears the board's "ROMs" (presumably system BIOS and VGA BIOS, at least) were loaded at start by the host, so it's possible these files exist somewhere in the OS media. I think the VGA on the board was essentially "copied" to the host video memory at runtime. All of these seems to indicate that the host and emulator could run concurrently.

PS: got enough SCSI working to enable the "rebuild" process to complete, and it's now crashing somewhere when attempting to boot the minimal rebuild environment.

Old Microsoft knowledge base article relating to Intergraph HardPC

It's been quite a while since I updated things here, mostly due to lack of substantial progress (not from lack of trying).

Things that have been tackled and at least partially implemented:

• support for the InterPro 6xxx family (deskside tower boxes, with 5 or 12 slot backplane)
• working toward a low level emulation of the keyboard (dumped thanks to Phil)
• most of the C400 instructions are working now (there's exactly zero documentation on these)
• CLIX is very, very close to booting (see screenshot below, not sure if I've shared before), but still stuck on something I've not yet been able to pinpoint
• located two people who have EDGE graphics boards (EDGE-1 and EDGE-2 Plus respectively), and started working on these

The EDGE (Extensible Display Geometry Engine) graphics hardware is interesting, largely built around between 1 and 3(!) TMS320C30 DSPs, and only available for the 6xxx family of systems. They are supposed to support between 8 and 24 bpp graphics modes, and perform hardware acceleration of both 2D and 3D operations. The boards have a block of static RAM as well as enough VRAM to handle double or even triple buffering in some cases. They also have mouse and keyboard ports, and seem to be used instead of the equivalent on the system board when installed. Interestingly, at least the EDGE-1 version seems to have been designed to support two concurrent users, each with their own mouse and keyboard, and presumably either 1 or 2 screens. I'm not sure yet if that functionality was available on the later models, but I thought it quite unusual and interesting.

The board below is the EDGE-2 Plus processor board, and can be paired with up to 4 framebuffer boards/screens I think. The 3 connectors on the top edge fit the backplane of the system, and I believe the connectors on the opposite edge are a kind of private bus handling the communication between the processor and framebuffer boards. It seems there's quite an extensive TMS320C30 diagnostic program, but I need to get the EPROMs from this board dumped before I'm likely to be able to get very far. At least the skeleton driver for these boards works as a simple framebuffer already, so it's possible to launch the FDM and experiment.

That's about it for now, but I have 3 main work fronts open - finish the keyboard LLE, get more of the EDGE graphics implemented, and figure out what's stopping CLIX from booting. The last would obviously be a very key milestone, but until now it has remained tantalizingly close yet out of reach.

Nifty. Booting more Unix variants is always good, even if this one isn't quite there yet

Good grief, that's a lot of PALs/GALs...