2 Dec
2014
2 Dec
'14
1:42 p.m.
I'm Roy Hirst, a hobbyist new to the list - today is my day 1 - and I
currently work for XKL LLC. I'd like to contribute where I can to this
and other conversations on cctalk.
XKL makes a line of optical networking systems, think 10x10GE in 1RU. A
system has two planes, a (1) traffic plane where client equipment (e.g.
"server") traffic gets aggregated to DWDM channels on a single optical
fiber, and (2) a management plane that, er, manages remote config,
real-time failover, OS field upgrades, etc. The planes are independent
and you can reboot the OS after an upgrade without affecting packet
traffic. The emphasis is on reliability, resilience, configurability,.
We have had DWDM systems running 5 years without incident using these
backplanes.
The TOAD-2 is looked back on here with affection, some of the original
architects and engineers are still around, and memory systems and
interfaces have been under continuous development and extension since
TOAD-2 days. I am not sure if you used to work here you would recognize
it now. Is still 36-bit,? Yes.
Roy
--
*Roy Hirst* | 425-556-5773 | 425-324-0941 cell
XKL LLC | 12020 113th Ave NE, Suite 100 | Kirkland, WA 98034 | USA
The information contained in this e-mail message may be privileged, confidential and
protected from disclosure.
If you are not the intended recipient, any dissemination, distribution or copying is
strictly prohibited.
If you think that you have received this e-mail message in error, please e-mail the
sender at the above e-mail address.
3 Dec
3 Dec
3:49 p.m.
Hi Roy
On 12/02/2014 07:42 PM, Roy Hirst wrote:
I'm Roy Hirst, a hobbyist new to the list - today
is my day 1 - and I
currently work for XKL LLC. I'd like to contribute where I can to this
and other conversations on cctalk.
Welcome to the list! Thanks for joining an shedding some light on the
36-bit/TOAD questions.
The TOAD-2 is looked back on here with affection, some of the original
architects and engineers are still around, and memory systems and
interfaces have been under continuous development and extension since
TOAD-2 days. I am not sure if you used to work here you would
recognize it now. Is still 36-bit,? Yes.
Perhaps you could elaborate on the compatibility between the TOAD
versions and other PDP-10 processors. Also what operating systems do
they run, both previous and current versions?
And, the question I suppose some of us have, where can I get one? (I
know there was only a handful TOAD-1s, but what about the TOAD-2, and is
there a TOAD-3?)
Regards,
Pontus.
4 Dec
4 Dec
2:25 a.m.
Pontus <pontus at update.uu.se> writes:
Perhaps you could elaborate on the compatibility
between the TOAD
versions and other PDP-10 processors. Also what operating systems do
they run, both previous and current versions?
I'll contribute the technical information I have. Happy to take
corrections or additions!
The TOAD-1 was XKL's first PDP-10 clone, created around 1994.
The CPU is called XKL-1, and is a model B with an extended virtual (30
bits) and physical address (33 bits) space compared to the KL10 model
B. It's clocked at 33 MHz, and has 128K of cache and 8K TLB entries.
The microcode control store has 8K 128-bit words.
Similar information for more 10s here:
http://pdp10.nocrew.org/cpu/processors.html
6:03 a.m.
On 2014-12-04 08:25, Lars Brinkhoff wrote:
Pontus <pontus at update.uu.se> writes:
As far as I can recall, the TOAD-1 is not compatible with the KL10B. I
don't remember what the differences were, but I understood that TOPS-20
had to be modified in an incompatible way to run on it. Which was
unfortunate, in that SC also made extensions, but they were more in line
with how the KL10B works. So we got two different branches when the
architecture was expanded from the KL10B. Unfortunately I don't remember
if it was the paging system or the I/O system where they diverged.
Someone who knows more details who can correct me on this?
Johnny
Perhaps you could elaborate on the compatibility
between the TOAD
versions and other PDP-10 processors. Also what operating systems do
they run, both previous and current versions?
I'll contribute the technical information I have. Happy to take
corrections or additions!
The TOAD-1 was XKL's first PDP-10 clone, created around 1994.
The CPU is called XKL-1, and is a model B with an extended virtual (30
bits) and physical address (33 bits) space compared to the KL10 model
B. It's clocked at 33 MHz, and has 128K of cache and 8K TLB entries.
The microcode control store has 8K 128-bit words.
Similar information for more 10s here:
http://pdp10.nocrew.org/cpu/processors.html
6:21 a.m.
Johnny Billquist <bqt at update.uu.se> writes:
As far as I can recall, the TOAD-1 is not compatible
with the
KL10B. I don't remember what the differences were, but I understood
that TOPS-20 had to be modified in an incompatible way to run on it.
Right, it's not entirely compatible in executive mode. But it is in
user mode (well, probably apart from some minor details).
3:38 p.m.
From: Lars Brinkhoff
Sent: Thursday, December 04, 2014 3:21 AM
Johnny Billquist <bqt at update.uu.se> writes:
> As far as I can recall, the TOAD-1 is not
compatible with the
> KL10B. I don't remember what the differences were, but I understood
> that TOPS-20 had to be modified in an incompatible way to run on it.
Right, it's not entirely compatible in executive
mode. But it is in
user mode (well, probably apart from some minor details).
The biggest detail in user mode is that the entire 30-bit address space
is available to user programs, where the Model B KL-10s provided a 22-bit
physical address and a 23-bit virtual (in TOPS-20 from v4 and in Tops-10
v7.04).
The user model is easily seen in TOPS-20 v7+ when invoking DDT: There is
a switch on the DDT command /USE-SECTION, the value of which is limited on
the KL to section 37 (octal), but can be up to 7776 on the Toad-1.
(Section 7777 is a magic value in the Toad architecture.)
The Toad-1 Architecture Reference Manual is available via anonymous FTP in
PDF, Postscript, and TeX DVI formats at Toad-1.LivingComputerMuseum.org,
in the directory <DOCUMENTATION.TOAD>. Remember to specify "tenex" or
"type l 8" for the PDF or DVI files.
Rich
Rich Alderson
Vintage Computing Sr. Systems Engineer
Living Computer Museum
2245 1st Avenue S
Seattle, WA 98134
mailto:RichA at LivingComputerMuseum.org
http://www.LivingComputerMuseum.org/
11:59 a.m.
Although here at XKL we don't still use TOAD-1s in the business, we do
have a few engineers here in R&D that are bit-level familiar with the
beast, even tho' the last living instance I believe is now at the Museum.
I'd need to stay well clear of anything XKL-proprietary, but I would be
happy to interview them and to put together a technical briefing on
stuff we can talk about. Would take me a week. Let me know if that would
help, and if there's any doc gaps that I might help fill. I'm new to
this forum, I can't tell what level of reverse engineering folk are
aiming at, are folk pursuing emulators, simulators, what? Does the
Museum provide some kind of a reverse-engineered BOM? I don't know where
you would actually buy a TOAD-1 (almost certainly not from us, I think,
but I will check) and our last one I suspect is the one proud in baby
blue on the Museum's video.
Roy
*Roy Hirst* | 425-556-5773 | 425-324-0941 cell
XKL LLC | 12020 113th Ave NE, Suite 100 | Kirkland, WA 98034 | USA
On 12/4/2014 3:03 AM, Johnny Billquist wrote:
On 2014-12-04 08:25, Lars Brinkhoff wrote:
The information contained in this e-mail message may be privileged, confidential and
protected from disclosure.
If you are not the intended recipient, any dissemination, distribution or copying is
strictly prohibited.
If you think that you have received this e-mail message in error, please e-mail the
sender at the above e-mail address.
Pontus <pontus at update.uu.se> writes:
As far as I can recall, the TOAD-1 is not compatible with the KL10B. I
don't remember what the differences were, but I understood that
TOPS-20 had to be modified in an incompatible way to run on it. Which
was unfortunate, in that SC also made extensions, but they were more
in line with how the KL10B works. So we got two different branches
when the architecture was expanded from the KL10B. Unfortunately I
don't remember if it was the paging system or the I/O system where
they diverged.
Someone who knows more details who can correct me on this?
Johnny Perhaps you could elaborate on the compatibility
between the TOAD
versions and other PDP-10 processors. Also what operating systems do
they run, both previous and current versions?
I'll contribute the technical information I have. Happy to take
corrections or additions!
The TOAD-1 was XKL's first PDP-10 clone, created around 1994.
The CPU is called XKL-1, and is a model B with an extended virtual (30
bits) and physical address (33 bits) space compared to the KL10 model
B. It's clocked at 33 MHz, and has 128K of cache and 8K TLB entries.
The microcode control store has 8K 128-bit words.
Similar information for more 10s here:
http://pdp10.nocrew.org/cpu/processors.html 
2:12 p.m.
On 12/4/14 8:59 AM, Roy Hirst wrote:
Although here at XKL we don't still use TOAD-1s in
the business, we do have a few engineers here in R&D that are bit-level familiar with
the beast, even tho' the last living instance I believe is now
at the Museum.
Museums. The Computer History Museum has one as well.
Wayne and Ralph are still there, I assume?
7:09 p.m.
Yes, Wayne and Ralph are definitely still here. Ralph was in my office
earlier wanting to know how he might help. A privilege to work with
these guys.
Roy
*Roy Hirst* | 425-556-5773 | 425-324-0941 cell
XKL LLC | 12020 113th Ave NE, Suite 100 | Kirkland, WA 98034 | USA
On 12/4/2014 11:12 AM, Al Kossow wrote:
On 12/4/14 8:59 AM, Roy Hirst wrote:
The information contained in this e-mail message may be privileged, confidential and
protected from disclosure.
If you are not the intended recipient, any dissemination, distribution or copying is
strictly prohibited.
If you think that you have received this e-mail message in error, please e-mail the
sender at the above e-mail address.
Although here at XKL we don't still use
TOAD-1s in the business, we
do have a few engineers here in R&D that are bit-level familiar with
the beast, even tho' the last living instance I believe is now
at the Museum.
Museums. The Computer History Museum has one as well.
Wayne and Ralph are still there, I assume?
3:48 p.m.
On 12/04/2014 05:59 PM, Roy Hirst wrote:
Although here at XKL we don't still use TOAD-1s in
the business, we do
have a few engineers here in R&D that are bit-level familiar with the
beast, even tho' the last living instance I believe is now at the Museum.
Besides the CHM one, Peter L?thberg also has one.
I'd need to stay well clear of anything
XKL-proprietary, but I would
be happy to interview them and to put together a technical briefing on
stuff we can talk about. Would take me a week. Let me know if that
would help, and if there's any doc gaps that I might help fill. I'm
new to this forum, I can't tell what level of reverse engineering folk
are aiming at, are folk pursuing emulators, simulators, what? Does the
Museum provide some kind of a reverse-engineered BOM?
I don't think there is any serious project going on, just general
curiosity. Please do correct me if I'm wrong.
I don't know where you would actually buy a TOAD-1
(almost certainly
not from us, I think, but I will check) and our last one I suspect is
the one proud in baby blue on the Museum's video.
I'm well aware of the TOAD-1 rarity, but what about the TOAD-2.
Regards,
Pontus.
3:52 p.m.
From: Roy Hirst
Sent: Thursday, December 04, 2014 9:00 AM
I don't know where you would actually buy a TOAD-1
(almost certainly not
from us, I think, but I will check) and our last one I suspect is the
one proud in baby blue on the Museum's video.
Hi, Roy,
The colors in the video must be off. It's definitely not baby blue, more
like a true sapphire.
Only four Toad-1 systems left XKL while we still offered them for sale:
Two were purchased by BT/Tymnet (one in private hands, one discarded by
people who did not know what it was), one went to Digital (now at CHM),
and one was purchased by Paul Allen for his Wired World project (now at
LCM). One more later went to a friend of the family known for collecting
PDP-10s.
Purchasing one would probably be difficult. :-)
Rich
(XKL sales support/customer support/TOPS-20 developer/technical writer/
systems administrator, 1993-2003)
Rich Alderson
Vintage Computing Sr. Systems Engineer
Living Computer Museum
2245 1st Avenue S
Seattle, WA 98134
mailto:RichA at LivingComputerMuseum.org
http://www.LivingComputerMuseum.org/
5:55 p.m.
That's very interesting to know. I had always wondered how many were made
and where they went.
Can anyone disclose the original list price of the TOAD-1? Or is that
off-limits? Just curious.
Best,
Sean
On Thu, Dec 4, 2014 at 3:52 PM, Rich Alderson <
RichA at livingcomputermuseum.org> wrote:
From: Roy Hirst
Sent: Thursday, December 04, 2014 9:00 AM
I don't know where you would actually buy a
TOAD-1 (almost certainly not
from us, I think, but I will check) and our last one I suspect is the
one proud in baby blue on the Museum's video.
Hi, Roy,
The colors in the video must be off. It's definitely not baby blue, more
like a true sapphire.
Only four Toad-1 systems left XKL while we still offered them for sale:
Two were purchased by BT/Tymnet (one in private hands, one discarded by
people who did not know what it was), one went to Digital (now at CHM),
and one was purchased by Paul Allen for his Wired World project (now at
LCM). One more later went to a friend of the family known for collecting
PDP-10s.
Purchasing one would probably be difficult. :-)
Rich
(XKL sales support/customer support/TOPS-20 developer/technical writer/
systems administrator, 1993-2003)
Rich Alderson
Vintage Computing Sr. Systems Engineer
Living Computer Museum
2245 1st Avenue S
Seattle, WA 98134
mailto:RichA at LivingComputerMuseum.org
http://www.LivingComputerMuseum.org/
6:18 p.m.
From: Sean Caron
Sent: Thursday, December 04, 2014 2:56 PM
That's very interesting to know. I had always
wondered how many were
made and where they went.
Oh, quite a few more were made. They were used in-house at XKL for
software development and CAD work as well as electronic mail and day-to-
day computing. They were finally fully retired a few years ago.
Can anyone disclose the original list price of the
TOAD-1? Or is that
off-limits? Just curious.
It was public information. List price for a minimal configuration was
$125,000. That comprised:
1 XKL-1 processor
1 XMG-1 memory (32MW @ 36b/W)
1 XNI-1 Ethernet interface (four 10baseT/AUI ports)
1 XRH-1 SCSI interface (four FASTWIDE differential ports)
1 DDS-2 DAT drive (internal)
1 4GB Seagate Barracuda SCSI drive (internal) with special microcode
to allow formatting at 2304 bytes/sector (= 512 words = 1 page)
The backplane had 7 slots, so additional memory and peripheral ports
could be added. We also OEM'd the M4data 9914 9 track tape drive with a
differential interface to allow companies to read archival data.
Eventually we began using DLT IV drives in-house for backups, having
learned the hard way that DAT was the wrong choice, and 9GB or 18GB disks.
We'd have OEM'd those two, if we'd still been in the business of selling
Toad-1 systems.
Rich
Rich Alderson
Vintage Computing Sr. Systems Engineer
Living Computer Museum
2245 1st Avenue S
Seattle, WA 98134
mailto:RichA at LivingComputerMuseum.org
http://www.LivingComputerMuseum.org/
7:13 p.m.
On 12/4/2014 3:55 PM, Sean Caron wrote:
That's very interesting to know. I had always
wondered how many were made
and where they went.
Can anyone disclose the original list price of the TOAD-1? Or is that
off-limits? Just curious.
Best,
Sean
Like high end audio equipment ...
If you have ask for the price, you can not afford it. :)
Ben.
3:54 p.m.
Roy Hirst <rhirst at xkl.com>
Although here at XKL we don't still use TOAD-1s in
the business, we do
have a few engineers here in R&D that are bit-level familiar with the
beast, even tho' the last living instance I believe is now at the Museum.
I'd need to stay well clear of anything XKL-proprietary, but I would be
happy to interview them and to put together a technical briefing on
stuff we can talk about. Would take me a week.
Personally, I'd be more interested in hearing more about the "TOAD-2"
controllers used in current XKL equipment, starting with:
Is the TOAD-2 a PDP-10 clone, as the TOAD-1 is?
Is it running TOPS-20 (or some subset)?
If not, is it a self-hosting system (is development for the platform
done on a TOAD-2) or is it an embedded processor, with binaries being
cross compiled on some other platform?
Thanks!
Phil (who once worked on DEC FORTRAN-10/20)
7:05 p.m.
Good questions, thanks. TOAD-2:
Not a PDP anything clone.
Embedded processor implemented on FPGA.
Software (and firmware) cross compiled on some other platform.
Instruction set and architecture optimized over several years for
real-time switching. Bad things happen very rarely but at 10x10Gbps when
bad things happen they can happen quickly.
Manages system environmental status (e.g. some of the optics, such as
filters, have tight dimensional tolerances and so can be temp sensitive)
and status and control of data plane laser transceivers. Data plane drop
rates must be < 1 bit in 10**16. Status and control is via a variety of
signal pins and buses all with custom hardware interfaces to this processor.
Reliability requirements are very high, systems must run unattended
(sometimes in inaccessible places) for 5-10 years 24x7 continuous. Laser
transceivers, fans, power supplies are all hot-swappable in the field,
but not the embedded processor.
In this embedded application there is a high value set on low power
consumption (especially since systems get data-center racked and so need
to be cooled). With this processor, you can run dual power supplies,
four fans, a manageability plane, 10x10G client laser transceivers plus
2 DWDM 10x10G 80km transport lasers, in one rack unit for the same power
as 2 domestic light bulbs. The proper 60-watt kind, not the new
supermarket funny ones.
Roy
*Roy Hirst* | 425-556-5773 | 425-324-0941 cell
XKL LLC | 12020 113th Ave NE, Suite 100 | Kirkland, WA 98034 | USA
On 12/4/2014 12:54 PM, Phil Budne wrote:
Roy Hirst <rhirst at xkl.com>
The information contained in this e-mail message may be privileged, confidential and
protected from disclosure.
If you are not the intended recipient, any dissemination, distribution or copying is
strictly prohibited.
If you think that you have received this e-mail message in error, please e-mail the
sender at the above e-mail address.
Although here at XKL we don't still use
TOAD-1s in the business, we do
have a few engineers here in R&D that are bit-level familiar with the
beast, even tho' the last living instance I believe is now at the Museum.
I'd need to stay well clear of anything XKL-proprietary, but I would be
happy to interview them and to put together a technical briefing on
stuff we can talk about. Would take me a week.
Personally, I'd be more
interested in hearing more about the "TOAD-2"
controllers used in current XKL equipment, starting with:
Is the TOAD-2 a PDP-10 clone, as the TOAD-1 is?
Is it running TOPS-20 (or some subset)?
If not, is it a self-hosting system (is development for the platform
done on a TOAD-2) or is it an embedded processor, with binaries being
cross compiled on some other platform?
Thanks!
Phil (who once worked on DEC FORTRAN-10/20)
7:47 p.m.
From: Phil Budne
Sent: Thursday, December 04, 2014 12:54 PM
Personally, I'd be more interested in hearing more
about the "TOAD-2"
controllers used in current XKL equipment, starting with:
I'm not sure I'd agree with the designation "controller".
Is the TOAD-2 a PDP-10 clone, as the TOAD-1 is?
The Toad-2 is an extended Toad-1 clone. It puts the entire Toad-1
multi-board system with improvements into a single FPGA.
Is it running TOPS-20 (or some subset)?
It's running a slightly later version of TOPS-20 than the Toad-1 which
employs the architectural enhancements of the Toad-2.
If not, is it a self-hosting system (is development
for the platform
done on a TOAD-2) or is it an embedded processor, with binaries being
cross compiled on some other platform?
Ralph has told us that (at least at one time) Toad-2 systems were being
used by (some of) the engineers as development platforms, so I would
call the Toad-2 self-hosting.
From what Roy Hirst has said, that may no longer be
true, but I have no
information on that point.
Rich
Rich Alderson
Vintage Computing Sr. Systems Engineer
Living Computer Museum
2245 1st Avenue S
Seattle, WA 98134
mailto:RichA at LivingComputerMuseum.org
http://www.LivingComputerMuseum.org/
3517
days inactive
3520
days old
16 comments
9 participants
participants (9)
-
aek＠bitsavers.org -
bfranchuk＠jetnet.ab.ca -
bqt＠update.uu.se -
lars＠nocrew.org -
phil＠ultimate.com -
pontus＠update.uu.se -
rhirst＠xkl.com -
RichA＠LivingComputerMuseum.org -
scaron＠umich.edu