From: mark@hubcap.clemson.edu (Mark Smotherman) Newsgroups: alt.folklore.computers,comp.sys.cdc Subject: CDC Flexible Processor? Date: 21 Mar 2002 13:56:51 -0500 Organization: Clemson University Lines: 19 Message-ID: NNTP-Posting-Host: localhost X-Trace: hubcap.clemson.edu 1016737012 1469 127.0.0.1 (21 Mar 2002 18:56:52 GMT) X-Complaints-To: usenet@hubcap.clemson.edu NNTP-Posting-Date: 21 Mar 2002 18:56:52 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!newsfeed.icl.net!newsfeed00.sul.t-online.de!t-online.de!deine.net!freenix!sunqbc.risq.qc.ca!news-ext.gatech.edu!hubcap.clemson.edu!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:104185 Does anyone have leads about the history and/or a detailed description of the CDC Flexible Processor and Advanced Flexible Processor (AFP)? [ What I found searching via Google said that the FP was developed as a horizontally-microprogrammed radar processor in the 1970's, it later developed into the AFP, and still later it was incorporated into the Cyberplus ring bus architecture. Each AFP instruction apparently could initiate operations on 15-18 functional units. Program memory was 4096 200-bit (or 240-bit with the FP option) words. ] Thanks. -- Mark Smotherman, Computer Science Dept., Clemson University, Clemson, SC http://www.cs.clemson.edu/~mark/homepage.html ###### From: kent@nettally.com (Kent Olsen) Newsgroups: alt.folklore.computers,comp.sys.cdc Subject: Re: CDC Flexible Processor? Date: 21 Mar 2002 13:54:04 -0800 Organization: http://groups.google.com/ Lines: 31 Message-ID: References: NNTP-Posting-Host: 63.90.83.178 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1016747645 20009 127.0.0.1 (21 Mar 2002 21:54:05 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 21 Mar 2002 21:54:05 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.imp.ch!news.imp.ch!news-spur1.maxwell.syr.edu!news.maxwell.syr.edu!newsfeed.stanford.edu!postnews1.google.com!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:104225 I was on the CYBERPLUS project for a while. It was a pretty cool concept in that it attached to a CYBER channel, loaded programs and data at channel speeds, ran like a scalded dog, and dump the results back to the host at channel speed. It never did support higher level languages -- all programming was in its assembly language. IIRC, UTexas developed an "optimizing assembler" that helped with instruction scheduling. I don't recall anything about 200+ bit words. I seem to recall that all memory and results were 16 bit. Kent mark@hubcap.clemson.edu (Mark Smotherman) wrote in message news:... > Does anyone have leads about the history and/or a detailed > description of the CDC Flexible Processor and Advanced > Flexible Processor (AFP)? > > [ What I found searching via Google said that the FP was > developed as a horizontally-microprogrammed radar processor > in the 1970's, it later developed into the AFP, and still > later it was incorporated into the Cyberplus ring bus > architecture. > > Each AFP instruction apparently could initiate operations > on 15-18 functional units. Program memory was 4096 200-bit > (or 240-bit with the FP option) words. ] > > Thanks. ###### From: "Lionel bening" Newsgroups: alt.folklore.computers,comp.sys.cdc References: Subject: Re: CDC Flexible Processor? Lines: 47 X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Message-ID: NNTP-Posting-Host: 12.237.22.3 X-Complaints-To: abuse@attbi.com X-Trace: rwcrnsc54 1016765392 12.237.22.3 (Fri, 22 Mar 2002 02:49:52 GMT) NNTP-Posting-Date: Fri, 22 Mar 2002 02:49:52 GMT Organization: AT&T Broadband Date: Fri, 22 Mar 2002 02:49:52 GMT Path: chonsp.franklin.ch!pfaff.ethz.ch!news-zh.switch.ch!news.mailgate.org!cyclone.bc.net!sunqbc.risq.qc.ca!nf3.bellglobal.com!wn1feed!worldnet.att.net!204.127.198.203!attbi_feed3!attbi.com!rwcrnsc54.POSTED!not-for-mail Xref: chonsp.franklin.ch alt.folklore.computers:104236 In the 1982-1983 days of CDC prosperity, the AFP project people were eager to get more software applications running on the AFP. So four of us from Arden Hills got trained on AFP assembly programming and started writing a logic simulator. Two programmers (the smart ones) realized the real degree of difficulty and quit the project immediately, and left CDC for Unisys shortly thereafter. The two of us that were left continued bravely onward. We found those inner loops fun to craft, but it took days to get each loop to execute its 16 VLIW instruction words correctly at the 50 MHz clock cycle rate. When we got them to work, they would be screaming at 800 MHZ instruction rate This was at a time when the Crays could run at a ~600 MHz peak instruction rate with chained vector instructions, at about 10X the cost of the AFP hardware. Alas, even though we could prove the concept in the inner loop speed, we did not have any of the user interface expected for a logic simulation application. The user interface that would have been easy to write in Pascal was too big a task to complete within an acceptable amount of time in the AFP assembly language. Because marketing had already hinted to customers about the forthcoming range of applications, stopping the AFP logic simulation project before completion took more justification than starting the project. The rumors were that the successful applications for the AFP were developed by the high-secrecy people - code-breakers and oil exploration. At the AFP programming training, over half the class were spooks. They would talk normally about their hobbies, movies and activities away from work but would say nothing about where they worked. Lionel