Thread View: comp.sys.transputer
2 messages
2 total messages
Started by Thomas E. Ragland
Sun, 06 May 1990 22:17
(none)
Author: Thomas E. Ragland
Date: Sun, 06 May 1990 22:17
Date: Sun, 06 May 1990 22:17
34 lines
2439 bytes
2439 bytes
Dear Netters, I'm currently seeking the assistance and advice from anyone that results answer several questions that I have. The problem area is the transputer and ientcy ability to be embedded into products relying on signal processing. For instance, based on the theory of concurrecy, we're trying to develop design approaches for a transputer architecture. Therefore, can someone in *netland* help resolve the following? The problem of interest is as follows, we would like to develop a specialized high speed parallel computer system for signal processing applications. However, complex systems based on the *transputer* can be constructed without truly realizing their operational behavior. Moreover, *transputer* technology, and its accompanying languages, and simulations do not, by themselves, provide sufficient assurances that important liveness and safety properties are fully satisfied, or that the best design has been developed. For expressing the above, we are considering several configurations based on current technology, namely, a Banyan network. Has anyone pursed signal processing/analysis on such an architecture? If so, were the -------------------------------------------------------------------------------- , formal analysis methods, algorithm design suggest feasible practical approaches to predicting the safety and liveness properties of *transputer* based systems. Finally, can anyoone in *netland* help expedite our research efforts by providing, references, citing(s), personal suggestions on developing *optimal* configurations based on the *transputer* for signal processing? Thanks in advance to all! ---------------------------------------------------------------------------------------------------------------------------------------------------------------- Thomas E. Ragland RTCL (Real-Time Computing Lab) Dept. of EECS The University of Michigan Ann Arbor, Michigan 48109-2122 Phone: (313) 763-4424 E-Mail: swank@dip.eecs.umich.edu Regards,
Re: Seeking a solution to an optimal Transupter configuration...
Author: Steven Ericsson Zenith
Date: Mon, 07 May 1990 18:10
Date: Mon, 07 May 1990 18:10
47 lines
3156 bytes
3156 bytes
In article <8987@hubcap.clemson.edu> swank@dip.eecs.umich.edu (Thomas E. Ragland) writes:
>help resolve the following? The problem of interest is as follows, we would
>like to develop a specialized high speed parallel computer system for signal
>processing applications. [... stuff deleted about using the transputer]
The transputer alone is unlikely to be a good choice for signal
processing applications, esp. if you are looking for high performance.
You shouldn't assume in this case that the parallelism in such a
system is ideally applied to a signal processing problem. However,
having said that, transputers make fine glue and combined with some
specialised signal processor can serve the purpose. INMOS, in fact,
produces several TRAMs which do exactly this. The VECTRAM (a high
speed T800 plus a ?? [I forget the name .. ZORAN? Does that sound
right?] signal processor) and I'm sure there must be a TRAM with
the INMOS A1xx series of cascadable signal processors attached.
Now, the person to speak to at INMOS on such matters is Hossein Yassie
(I'm sure he won't mind my giving his name out and singing his praises
here). He understands all the issues mentioned in the original posting
and since he was the principle architect behind the A1xx series of
signal processors and responsible for the VECTRAM, he's a better
person than I to answer any detailed questions on this issue. He is
currently running the INMOS IQ Systems business but maintains a high
profile technical role. Sorry, email to INMOS Bristol,UK is not a
viable means of communication from the USA, but you could try
"isnet!bra!hossein@inmos.com". The land mail address is:
INMOS Limited,
1000 Aztec West,
Almondsbury, BRISTOL, UK.
Blissfully I realise I now forget the phone number! :-)
Steven.
. .
Steven Ericsson Zenith * email: zenith@cs.yale.edu
Department of Computer Science | voice: (203) 432 1278
Yale University 51 Prospect Street New Haven CT 06520 USA.
"All can know beauty as beauty only because there is ugliness"
.
Steven Ericsson Zenith * email: zenith@cs.yale.edu
Department of Computer Science | voice: (203) 432 1278
Yale University 51 Prospect Street New Haven CT 06520 USA.
"All can know beauty as beauty only because there is ugliness"
Thread Navigation
This is a paginated view of messages in the thread with full content displayed inline.
Messages are displayed in chronological order, with the original post highlighted in green.
Use pagination controls to navigate through all messages in large threads.
Back to All Threads