Article 4647 of alt.sys.pdp10: Path: news3.best.com!news2.best.com!newshub.northeast.verio.net!logbridge.uoregon.edu!news.u.washington.edu!Tomobiki-Cho.CAC.Washington.EDU!mrc From: Mark Crispin Newsgroups: alt.sys.pdp10 Subject: Re: The True Faith Date: Wed, 30 Dec 1998 12:01:03 -0800 Organization: Networks & Distributed Computing Lines: 20 Message-ID: References: <36794982.265795064@netnews.worldnet.att.net> <3681A3D8.A6192E70@MA.UltraNet.Com> <76d1h6$mcn$1@shell3.ba.best.com> <76d5ti$nqh$2@strato.ultra.net> NNTP-Posting-Host: tomobiki-cho.cac.washington.edu Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Trace: nntp1.u.washington.edu 915048065 42112 (None) 140.142.17.35 X-Complaints-To: help@cac.washington.edu NNTP-Posting-User: nark In-Reply-To: <76d5ti$nqh$2@strato.ultra.net> Xref: news3.best.com alt.sys.pdp10:4647 On Wed, 30 Dec 1998 jmfbahciv@aol.com wrote: > Why were those KL extend instructions implemented? I heard that it was for the COBOL compiler. The only EXTEND instruction that was routinely used by normal people was MOVSLJ, which was a faster string copy than a ILDB/IDPB/SOJG loop and even some of the hairy byte blt algorithms. Even then, MOVSLJ was mostly used by folks who wrote extended addressing programs. I guess that XBLT was probably also useful for someone, but I don't recall ever having to use it. If you're doing that sort of thing, it's probably much better just to change the page mapping. I mean, how often do you want to make a private copy of data from one section into another? -- Mark -- * RCW 19.149 notice: This email address is located in Washington State. * * Unsolicited commercial email may be billed $500 per message. * Science does not emerge from voting, party politics, or public debate. Article 4648 of alt.sys.pdp10: Path: news3.best.com!nntp1.ba.best.com!not-for-mail From: inwap@best.com (Smith and O'Halloran) Newsgroups: alt.sys.pdp10 Subject: Re: The True Faith Date: 30 Dec 1998 15:52:06 -0800 Organization: Chez Inwap Message-ID: <76eeb6$gn7$1@shell3.ba.best.com> References: <36794982.265795064@netnews.worldnet.att.net> <3688DE37.6445EFFD@s054.aone.net.au> <76d1h6$mcn$1@shell3.ba.best.com> <76d5ti$nqh$2@strato.ultra.net> Lines: 32 NNTP-Posting-Host: shell3.ba.best.com X-Trace: nntp1.ba.best.com 915061936 9736 inwap@206.184.139.134 Xref: news3.best.com alt.sys.pdp10:4648 In article <76d5ti$nqh$2@strato.ultra.net>, wrote: >Why were those KL extend instructions implemented? One reason was so that the COBOL compiler with BIS (Business Instruction Set) could do BCD arithmetic directly. Under the EXTEND opcode were: 1) "G" format floating point numbers (2.8E309 instead of 1.7E38) 2) string move and compare 3) XBLT 4) convert binary to string and back 5) the godawful EDIT instruction. The four most common G-float instructions were added where there were holes in the opcode table (GFAD=102, GFSB=103, GFMP=106, GFCV=107). But the opcode to convert G-float to F, D, 36-bit int and 72-bit int had to go under EXTEND because there were no more opcodes available. XBLT was very useful. But the rest of the extended instructions were useful mainly for the people who wrote the COBOL compiler. Does anyone have benchmarks showing how much faster a KL with microcode version 271 or later and COBOL-74/BIS was compared to KL with the old microcode? I don't think it was that impressive. IMHO the BIS wasted precious microcode space. I remember seeing one comment in the release notes saying "That's it folks - the last available word of microcode has been used. There will be no further improvements." -Joe -- INWAP.COM is Joe Smith, Sally Smith and our cat Murdock. (The O'Hallorans and their cats moved to http://www.tyedye.org/ Nov-98.) See http://www.inwap.com/ for PDP-10, "ReBoot", "Shadow Raiders"/"War Planets"