rtVAX 300.

I don't know about other rtVAXen, if any, but the KA620 I know reasonably well - I once built the kernel to manage the KA620 in a dual-processor KA630/KA620 system (and the OS hackery to speak to it Based on that: the KA620 is just like the KA630, except that P0 and P1 page tables live in physical space instead of kernel virtual space. That is, P0BR and P1BR must be computed based on the page tables' physical addresses rather than their kernel virtual addresses. (This means they need to be physically contiguous rather than virtually contiguous, which is one potential issue.) Of course, if you're doing something that never uses P0/P1 space - stay entirely in kernel mode, or perhaps never even turn on MAPEN at all - you might be able to totally ignore this. But it won't run VMS, which I gather is what DEC cared about at the time. /~\ The ASCII Mouse \ / Ribbon Campaign X Against HTML mouse at rodents-montreal.org / \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B