strange number corruption on pdp11/34

[John Wilson]

On Tue, Jan 27, 2015 at 11:06:35PM -0500, Jacob Ritorto wrote:
>On Tue, Jan 27, 2015 at 10:28 PM, js at cimmeri.com <js at cimmeri.com> wrote:
>>
>> (3) 11/34A boards?
>
>
>Right, 8266, 8265, 8267.  I assume the 8267 is what makes it an "-a" with
>floating point, right?

The *ability* to accept the FPP (and/or cache -- M8268 I think?) is what
makes it an -A.  And you need the right over-the-top connector (hazy memory
says H882 or H8822 for either or both option but I may have made those up).
Better to reduce the variables and leave the FPP off for now.

The fact that you can get so far with such a fundamental bug seems insane.
I'd almost start wondering about your terminal (it's insane, but less
insane), if you weren't copying/pasting all those all-zeros numbers.  XXDP
stuff would never depend on DIV or ASH unless that's what it's testing --
its whole purpose is to limp along on partially broken hardware.  Insanity!

John Wilson
D Bit