• Please review our updated Terms and Rules here

Mercury Computer Systems VAX Clone / Co Processor?

K

klapperp

Experienced Member
Joined
Dec 14, 2021
Messages
84
Location
Austria
Hi there,

Is there anybody who is familiar with this equipment?

From time to time there is offered such a board at ebay:

s-l1600-11.jpg

Currently here:

www.ebay.com

Mercury Computer Systems AP3232QB QBUS - Warranty | eBay

We have refurbished the part. VAX 6000, VAX-11/730. ABOUT KEYWAYS.
www.ebay.com

The seller wants a lot of money for it and I ask myself what it really is? The Weitek chip reminds me more on Sun then on Digital but you never know.

When you search at the site keyways.com you get a few more results for "Mercury"

MERCURY COMPUTER SYSTEMS 56-000900
MERCURY COMPUTER SYSTEMS 56-001000
MERCURY COMPUTER SYSTEMS 56-001100
MERCURY COMPUTER SYSTEMS AP3232QB
MERCURY COMPUTER SYSTEMS AP323QB
MERCURY COMPUTER SYSTEMS BM3200QB
MERCURY COMPUTER SYSTEMS CP3200QB
MERCURY COMPUTER SYSTEMS XM3200QB

I am wery interested in getting more information about this.

Any information is welcome.

Thanks in advance,

// Peter
 
Hi Al!

Thanks for your fast reply!
Is it a standalone system or an enhancement for a VAX and does this board alone make any sense?
Is there any information like a manual available anywhere?

// Peter
 
I have one of those boards. I don't know what the Weitek chips are, but the big AMD chips under the black heat sink are Am29325s, which are very nice pipelineable FPU chips. We built a supercomputer around a bunch of those at Princeton University in the 1980s, for a NASA contract.

The board is physically Qbus, but if you look at the card-edge connectors, it doesn't actually appear on the Qbus. I believe these were intended to be used in a group, connected together with ribbon cables to those connectors at the top, then over to some sort of a bridge board that also connects to the host computer.

-Dave
 
Hi Dave,

Wow, do you also have any documanetation about it? I am asking all those questions because I acquired probably a unit of whatever it was, cosnsiting of a bunch of boards, together with a QBUS chassis which I primarily wanted:

s-l1600-1.jpg

And now the whole thing is waiting in my storage (the left one on the picture):

20200901_203805 (576x1024).jpg

The more I am thinking, the more it begins to make sense for me to bring this unit back to life, because it seems to be an interesting piece. Maybe we can do this together and at the end we have two of these units alive?

// Peter
 
It would be amazing to get that running again, but unfortunately I have no documentation or software. I only have one (actually I think two) of the array processor boards. I think the big key would be to find either software (presumably libraries for the host computer) or at least register-level programming documentation.

-Dave
 
I have got some documents from a very nice person at Mercury Systems after I was begging a lot ;-) Maybe you find somebody else who was involved in the project for NASA at the Princeton University in the 1980s and who is also interested to bring this equipmet back to life. I can share the documents with you but better within a PM, just let me know.

// Peter
 
Wow, it's great that you were able to get some info from Mercury!
However, you misunderstand, we did not use that Mercury system at Princeton. We built that supercomputer completely from scratch, and coincidentally, Mercury used one of the same chips that we did (the Am29325), that's all.

-Dave
 
mcguire said:
I have one of those boards. I don't know what the Weitek chips are...
Click to expand...

WTL 1066 - High speed 32-bit x 32-word register file with three read ports and three write ports. 240 million bytes-per-second total data transfer rate (WTL 1066-10). Exponent ALU and 7-bit look-up ROM. First approximation for reciprocal device and square root sequence.

WTL-1066_Block-Diagram.png
 
gslick said:
WTL 1066 - High speed 32-bit x 32-word register file with three read ports and three write ports. 240 million bytes-per-second total data transfer rate (WTL 1066-10). Exponent ALU and 7-bit look-up ROM. First approximation for reciprocal device and square root sequence.
Click to expand...

Wow, zoom! That chip looks like quite a lot of fun.

-Dave
 
To be precise - we have a little bit of material:

doc-sample-1.jpg
doc-sample-2.jpg
doc-sample-3.jpgdoc-sample-4.jpg

Don't know if its enough to compute something on this equipment but definitely enough to get a clearer picture what we can probably do with it.

// Peter
 
If we will be nice enough and can present an interesting project, maybe. At the end of the day it could also be intersting for them, to support such a project.

// Peter
 
klapperp said:
If we will be nice enough and can present an interesting project, maybe. At the end of the day it could also be intersting for them, to support such a project.
Click to expand...

Well, most corporations won't bother unless they see a large profit opportunity, but you never know, you might get lucky.

-Dave
 
Just save them please, if you didn't have already done. It's always good to have more than one source ;-)
I let you know when I have time to do any further ...
// Peter
 
You must log in or register to reply here.
Back
Top