27 Sep
2012
27 Sep
'12
2:45 p.m.
On Thu, Sep 27, 2012 at 2:51 PM, Dave McGuire <mcguire at neurotica.com> wrote:
On 09/27/2012 01:39 PM, Rich Alderson wrote:
>>
>> Well, XKL's TOAD-1 is "an FPGA emulation", strictly speaking...
>
> Not, strictly speaking, true. The FPGAs + AMD 29xx sequencer in the XKL-1
> CPU implement a microcode engine, and the extended PDP-10 architecture is
> then implemented in microcode just as in the KL-10 and KS-10 processors.
Interesting to know. There's a lot I could learn about each and every
implementation.
> The later XKL-2 processor, implemented in a single
FPGA, is much closer
> to what Ethan means, I would think, although still microcoded.
Yes. That would count, but I doubt I'll ever land any XKL hardware.
> Perhaps Ethan has in mind something more like what
Neil Franklin attempted
> a decade ago, or "dgcx" more recently.
That is more what I was thinking of - a slab of silicon that (above
the microcode level) runs the PDP-10 instruction set native, and
allows for some modern or semi-modern disks to be attached (I really
don't care if it's SCSI or PATA or SATA or SAS or FC-AL or whatever -
just so I can mount something at least as large an RP06 or RP07) _and_
Ethernet (that acts at the register level like "real" hardware -
NI-10? NI-20? Is that the right nomenclature?)
That's a very blurry distinction, and is a path
directly into quicksand on
this list. ;) Where DOES one draw the line between hardware and firmware?
Several PDP-11s were microcoded. VAXen were routinely microcoded. I
don't plan to implement a KA-10 in ECL, so as long as I turn it on and
it doesn't have to boot another OS on some Intel-compatible processor
then fire off an app written in C to implement the PDP-10 instruction
set (something I'm already doing just fine with klh10), it's gotta be
closer than what I have now.
-ethan