LC ITS

[larsbrinkhoff]

LC ITS is to be a new KS10 configuration:

• RH11 + eight RP06s.
• 512K memory.
• TM03 + one TU45.
• DZ11.
• Chaosnet.
• Ethernet? (Depends on @atsampson.)

CC @rmaldersoniii.

[rmaldersoniii]

The KS10 for which this is destined has a DZ11, a Chaosnet board, and an Ethernet card which Bawden said was someone's project for a while (drivers in a personal directory).

[larsbrinkhoff]

Is the Ethernet card an Interlan NI1010? We may be able to get that to work, but probably at a later time.

[atsampson]

Doesn't that machine have an ITS name (and CONFIG entry) already? :-)

[larsbrinkhoff]

The lars/lc-its is a work in progress. It's still only using one RP06.

[rmaldersoniii]

Of course the machine had a name, but that one is well known and has built-in expectations. We want to start fresh, with more disks hosting more software and the possibility of allowing public accounts. I thought I had sent the LC config to Lars several months ago. If not, I apologize; I last touched the build in January 2018, then got buried in a different OS and architecture altogether.

[larsbrinkhoff]

The microcode RAM file is built fresh from ITS microcode sources. It's not binary identical to the one that came from here: http://www.cosmic.com/u/mirian/its/files/minsys.tape

@blipvert, do you know where the .; RAM RAM file on the minsys tape came from?

[larsbrinkhoff]

I have eight RP06 images built for LC ITS. I used the KLH10 emulator for this, since it supports Chaosnet. Unfortunately, it only has a stub device for DZ11 so I haven't tested that.

I haven't considered the file format. I believe the disk emulator uses the SIMH format, so I think I need to rebuild with KLH10 using that. Then I can cross check the images with SIMH, but I'm quite sure it'll complain about CH11 missing.

[larsbrinkhoff]

It just occurred to me, we should probably increase NUDSL to have more user directories available. The current value of 500. is large for an ITS, but I guess LC will be even larger in terms of disk space.

[rmaldersoniii]

Yes, the MDE uses SimH format for disk images, but KLH10 is agnostic about disk images. If instead of declaring a disk "DBD9", you declare it "DLW8", you get a SimH-compatible disk image and don't have to worry about CH11 not existing on the SimH KS10 emulation.

[larsbrinkhoff]

I'm using DLW8 now, and verified that SIMH likes this format. The disk image will boot to DSKDMP.

I see the image files vary in size. Does MDE need them padded with zeroes?

[larsbrinkhoff]

Possible issues:

• Microcode wrong?
• DZ11 boards, are they working?
• Can ITS even be run without DZ11? KLH10 implies not.

[larsbrinkhoff]

I think we can claim the microcode is verified now. The version running is freshly built from source code.

[larsbrinkhoff]

• CH11 board, is it ok?
• NI1010 board. ITS support hasn't been finished. But if it has been sitting there all along, maybe there's no harm in leaving it there.

[rmaldersoniii]

Because LC uses an LCM+L MDE rather than DEC spinning rust, it may be that the KS can handle RP07 images (where real RP07s are too fast for the KS-10's Unibus). At some point we should test this out.

[larsbrinkhoff]

I understand the MDE doesn't emulate a tape drive. That is unfortunate, because backing up to tape is a convenient way to move files from one system to another. Or the same system with a different disk configuration.

[rmaldersoniii]

There is a variant, the MTE, which works as 8 TM78 controllers with a single drive each (because subunits are difficult to implement in the Xilinx). The KS-10 does not support the TM78, only the TM03/TM02 style controller. One of the engineers is working on making the MTE talk TM03, but it's not high on his priority list.

[larsbrinkhoff]

I believe RP07 support was added at the very last minute and never tested at MIT and hence quite possibly never used on a real KS10 running ITS. I don't know if anyone ran it with KLH10 or SIMH; SV and UP both use RP06.

[rmaldersoniii]

I used an RP07 early in my experiments with your Github project, under KLH10 and SimH. It worked OK as I recall for a single-disk system, but multiple RP07s got into trouble very quickly. Since this was not a priority at the museum, I did not look into what was actually failing. Seems like something for a younger hacker to think about.

[larsbrinkhoff]

I'm breeding a young hacker, but he's only five.

[larsbrinkhoff]

CC @SMJ. I suppose most issues have been resolved now.