1GHz PDP-11

From: Jerome Fine <jhfine_at_idirect.com>
Date: Thu Dec 13 09:03:45 2001

>Doc wrote:

> On Thu, 13 Dec 2001, Dave McGuire wrote:
> > I don't recall who at the moment...but there's some company
> > somewhere who makes a PCI-Qbus bridge, and maybe a Unibus one as
> > well. The idea is to use a pdp11 emulator on a crappy PeeCee and be
> > able to connect your DEC peripherals to it.
> Here's a list, Unibus & Qbus PCI adapters:
> http://www.dbit.com/adapters.html

Jerome Fine replies:

Just a few words about the above adapters. Note that this information
is second hand since I have not used them myself, only been in direct
contact with the actual users.

V2.2 of Ersatz-11 (E11) had some problems with use in a multi-processor
environment when the TLC adapters were being used. From the information
about the problem, it was only going to cause a difficulty if there was more
than one PDP-11 being emulated (Multi-Processor) at the same time. The
reason the difficulty was not causing a problem in V2.2 was that no one was
attempting a multi-processor configuration like a PDP-11/74 system - not that
E11 even supported such a configuration in V2.2 of E11. The actual reason
for the problem was that the internal code within V2.2 of E11 did not correctly
do a Read/Modify/Write cycle for an instruction such as:
   Bis #100,_at_#CSR
rather, first a Read, then internally the Modify and finally a full Write cycle.
This could have allowed a second CPU to also begin the same sequence
at almost the identical time and thereby result in a timing problem.

With V3.0 of E11, this issue has been corrected. While the Unibus users of
E11 have not found a difficulty with the software/hardware being used with
a BCI-2004 adapter, the Qbus users with a BCI-2104 adapter had found
there is a problem which causes the incorrect execution of the above
instruction with a Qbus board in a BA23 box and V3.0 of E11 using the
BCI-2104 adapter. On the other hand, using V2.2 of E11 runs everything
in a satisfactory manner. At the present time, there is a suspicion that the
problem could be with the timing for the Qbus board which obviously
functions correctly in that BA23 box when used with a real DEC PDP-11/73
CPU, but not when V3.0 of E11 is used with the adapter. Or the firmware
and hardware associated with the BCI-2104 could be the problem. Note
that if the instruction is changed to a:
    Mov #100,_at_#CSR
then there is only a Write cycle and everything is again OK even with V3.0 of E11.

Also, realize that the adapter is almost as expensive and the first license for
the Commercial Version of Ersatz-11, the cost being about $ US 2500 and
$ US 3000 ($ US 4000 for the Linux version - although I understand the
Linux version of E11 still does not support the adapters), respectively.

On the other hand, no DEC PDP-11 today is able to run at the speed of even
a Pentium III system. And even the PDP-11/93 CPU cards are almost as
much today as a Commercial E11 license. But a Mentec Qbus CPU or a
QED CPU card that attempts to approach the speed of E11 on a Pentium IV
is probably far more expensive than the combined cost of E11 and the PC.
Of course, if a company requires the extra speed and they have hardware
which can't be emulated (which means there can be problems with the
interaction between the adapters and the emulator, then a full Qbus only
system is probably best.

As for the person who mentions the Charon-11 emulator,

>Ian Koller wrote:

> Dave,
> > I don't recall who at the moment...but there's some company
> > somewhere who makes a PCI-Qbus bridge
> Mentec offers those, as well as a PDP-11 emulator.
> http://www.mentec.com/mentecinc/emulation.htm
> http://www.mentec.com/mentecinc/PCIQ_Adapter.htm

I understand that the fellows who use E11 and the Unibus hardware have
attempted to use the Charon-11 as well and found it to be very unsatisfactory.

By the way, Charon-11 is NOT a Mentec product, as far as I know, Mentec
just pretends it is and sell the Charon-11 emulator as if it is a Mentec product.
My guess is that Mentec did not want to have Ersatz-11 as their "official"
emulator offering for the kind of reasons that companies have when the
decision is made to offer a product which is not as technically capable
as some other product, i.e. I have been told that E11 is much better than

Sincerely yours,

Jerome Fine
Received on Thu Dec 13 2001 - 09:03:45 GMT

This archive was generated by hypermail 2.3.0 : Fri Oct 10 2014 - 23:33:38 BST