Info-IBMPC@USC-ISIB.ARPA (Info-IBMPC Digest) (03/10/86)
Info-IBMPC Digest Sunday, 9 March 1986 Volume 5 : Issue 31 This Week's Editor: Phyllis O'Neil Today's Topics: IP/TCP NRC & MIT DBase III, printers, and LPTX (5 messages) Microsoft C V 3.0 and Phoenix Linker PLINK86 V 1.30 MS C 3.0 Large model struct assign code generation BUG Optimizer for DeSmet C compiler `C` sources & MS-DOS implementation of MIT teco editor Wish list for Microsoft C compiler Emacs-like editor (2 messages) Queuing System Simulation Software Query FFT Boards, Frame-Grabber Hardware FSCRC contract... Incremental File Backup Utilities Turbo Prolog DOS Termination Priam Drives on AT, Wait State on AT Limits on files open: concurrent and lifetime. LSPELL added to Library Dirlib Package added to Library Kermit 2.28 jrd/5b + Turbo Lightning /G0 and /G1 benchmarks NANSI (2 messages) Today's Queries: Turbo's 64K Limitation Query Compaq Deskpro and NEC V30 Query Qubie Modem Problem Lotus 1-2-3 Version 1 vs. Version 2 Query GEM Draw Problem ICOM R71A Radio Receiver to PC Query CMS Hard Disk, IBM Proprinter, Write & Spell Query Ctags for MS-DOS wanted Character ROMS Query Serial printer on PCjr Query Real-time PC operating systems query Serial ports query Graphics query Clearing the screen query Alternate Character Sets Query Program chaining query Wang PC "Open Architecture" package query ---------------------------------------------------------------------- Date: March 9, 1986 From: Richard Gillmann (Gillmann@ISIB) To: Info-IBMPC@ISIB Subject: New Editor We would like to welcome Phyllis O'Neil of the Aerospace Corporation as a new editor of Info-IBMPC Digest. Phyllis can be reached at oneil@aerospace.arpa. Richard ------------------------------ Date: Fri, 7 Mar 86 01:25:43 est From: romkey@BORAX.LCS.MIT.EDU (John Romkey) To: info-ibmpc@usc-isib.arpa Subject: IP/TCP NRC & MIT Well, it seems about time that I posted something as a followup to Billy Brackenridge's message about NRC's TCP/IP and my code. I've just been incredibly busy the last couple of weeks; keeping a small company going is very time consuming. I've never used NRC's code, but a number of the problems Billy pointed out with it (terminal emulation, host name resolution) have been fixed for a long time in the MIT PC/IP code. Now I'm taking the MIT PC/IP code and selling it myself, after having spent several years developing it. A couple of other companies seem to be selling an older version of the MIT code: Spartacus and Wollongong, for those who are interested in knowing. The new implementation of PC/IP is going to correct (hopefully) many of the problems with it that I've seen over the past few years but haven't, for one reason or another, been able to correct. For instance, there are a number of new applications: FTP, rlogin, rsh, rexec, rcp, rwho. The terminal emulator has been 'fixed' so that it will work on a greater variety of display hardware, but still be fairly fast (PC/IP's had lots of problems on non-IBM displays). And it's now written in Microsoft C - it compiles faster on my AT than it did at MIT on a VAX 11/750. There will also be a set of programming libraries available with both the native PC/IP programming interface, and the demented 4.2BSD programming interface in them so that people can write their own applications. The new code is marketed by FTP Software, Inc., of Cambridge, MA. Our address is: FTP Software, Inc. PO Box 150 Kendall Square Branch Boston, MA 02142 (617) 497-5066 Odd, the Post Office Box is actually located in Cambridge, but its address is in Boston... That's all I'll say for now about the product; I don't want to get into advertising on the net. If anyone has questions about it, feel free to ask me. Billy also suggested that we use INFO-IBMPC as a forum for discussing TCP/IP implementations. Here's a list of all the available implementations that I know of for the PC. A * means it's not commercially supported: Excelan (requires their hardware) FTP Software PC/TCP *MIT PC/IP NRC Fusion Spartacus KNET/PC Wollongong WIN/PC CMU and Stanford have done some additional work to MIT PC/IP that they are making available to others, at least to academic sites. There are others who have done work too, I'm just not sure that they're distributing it. So there seem to be about three strains of TCP/IP for the PC here: Excelan's, which I know little about save that it's mostly on their outboard processor; MIT's; and NRC's. Anybody know of any other TCP/IPs for the PC? Want to describe what's in these packages, your experiences with them? I'd also be interested in hearing what some people think would be useful new protocols to implement. Most of the current TCP/IP implementations implement the 'standard DARPA protocol suite': telnet, FTP, *maybe* SMTP, and some random protocols. The problem is, these protocols were all designed for use by somewhat sophisticated users on mainframes, and they're fine for that. But sharing files between PC's using FTP is simply gross. So we have network filesystems nowadays (yes, the reasoning does *not* quite follow...) What other sorts of applications would be useful? I could see having a dictionary server on a CD ROM based machine. It would be nice to have some sophisticated printer spooling and user database (so that you can get some info on *who* is using the PC - what's their favorite color?) protocols built on top of TCP/IP. Dave Clark once suggested a protocol that took me a while to digest. He suggested implementing a program that allows you to basically telnet to a PC from another machine - and get to use to see the graphics display on your local bitmapped display. That way you could access some of the more interesting software on the PC from your bigger and better personal workstation without having to rewrite it for the workstation. I finally decided it was a somewhat neat idea. Perhaps this kind of discussion would not be appropriate for INFO-IBMPC. I don't think it quite fits into the way the TCP-IP list works, either. If so, we could always start up a new list elsewhere if enough people were interested. John Romkey romkey@borax.lcs.mit.edu President, FTP Software Disclaimer? What disclaimer? I'm hardly unbiased... ------------------------------ To: Info-IBMPC@usc-isib Subject: Problems with lptx.asm Date: 06 Mar 86 17:19:20 EST (Thu) From: opus@athena.MIT.EDU I just recently downloaded the LPTX.ASM program from the library, and tried assembling it with Microsoft MASM 3.0. The resulting .COM program refuses to work correctly. The first time I tried to run the program, it printed up the banner, and then froze the machine, forcing a cold boot. I experimented in debug, and successfully managed to trace my way through it without a problem. This led me to believe that the program was interfering with one of my resident programs/drivers. On a hunch, I removed NANSI.SYS from my config.sys file, and re-booted. Wonder of wonders, it worked! Now, question number 1 is: Which program is giving me the problems, LPTX or NANSI? Unfortunately, this was not my only problem with LPTX. After running it, and redirecting lpt1: to the default file, I tried typing DIR > PRN, to test the program. The machine halted, and I had to warm boot. I examined the LPTX1.LST in the root directory, and noted that it contained 512 bytes. I tried printing a smaller file, and the program worked correctly. Now my question is: why doesn't LPTX work with data larger than 512 bytes? Thanks in advance for the help. David Jedlinsky opus@athena.mit.edu [Are you running DOS 3.1? -rag] ------------------------------ Date: Thu 6 Mar 86 20:37:16-PST From: David John Buerger <D.Buerger%SCU%PANDA@SUMEX-AIM.ARPA> Subject: Re: LPTX To: INFO-IBMPC@USC-ISIB.ARPA In response to Steven Mahan's request for some way to output from dBase III to a disk file, I've successfully used a public domain software program called LPTX. This program allows one to define parallel output to any disk file of your choice. You may then detoggle this command and continue output to your printer. Unfortunately, it only seems to work under DOS 3.1; I tried repeatedly to make it work with DOS 2.1 and it didn't work. David J. Buerger Director, PC Center Santa Clara University Dave%SCU%Panda@SUMEX-AIM.ARPA ------------------------------ Date: Friday, 7 March 1986 13:46:05 EST From: Joe.Newcomer@a.sei.cmu.edu To: info-ibmpc@usc-isib.arpa Subject: DBASE III output-->printer DBASE III works under some fairly serious delusions, not the least of which is that all PCs have printers and "print" means "send to the physical printer device". This is of course worsened by the inability of DOS to do even the most primitive redirection of logical devices. My solution in DBASE III is to totally ingore anything in the manual with the word "print" attached, and use the "set alternate" mechanism to direct output to a file. The use of "set alternate" is not obvious, but once you know this is the mechanism, HOW to use it is fairly straightforward. set talk off is a good mode to work in. Using the set alternate method, anything printed on the screen goes to a file, so most of the reports can be redirected this way. joe ------------------------------ Date: Fri, 07 Mar 86 13:09:14 EST From: ST602556%BROWNVM.BITNET@WISCVM.WISC.EDU To: Info-IBMPC@USC-ISIB.ARPA Subject: dBASE III output In reply to the query about changing COM1 to be a file in order to spool dBASE output to the disk instead of the printer: Every dBASE output command that I know of provides a method of sending the output to a disk file rather than the printer. See the SET ALTERNATE and SET DEVICE commands and various options of the LIST and DISPLAY commands. Eric Wolf ST602556 @ BROWNVM.BITNET ------------------------------ Date: 9 Mar 1986 15:54-EST Subject: Re Capturing Print Output from Dbase III From: ABN.ISCAMS@USC-ISID.ARPA To: steve@NCSC.ARPA Cc: info-ibmpc@USC-ISIB.ARPA Steve, I routinely use LPTX (LPTX.ASM, available via anonymous ftp from INFO-IBMPC archives at USC-ISIB.ARPA). I do NOT work with Dbase II or III, so cannot guarantee LPTX will work with that program .. but it appears to work with everything else I've tried. Price is right too. Regards, David Kirschbaum Toad Hall ABN.ISCAMS@USC-ISID ------------------------------ Date: 9 MAR 86 00:00-AST From: IU%DACTH51.BITNET@WISCVM.WISC.EDU To: INFO-IBMPC@USC-ISIB.ARPA Subject: Microsoft C V 3.0 and Phoenix Linker PLINK86 V 1.30 Two friends of mine are working on a comprehensive word processor using the above C compiler from Microsoft. Due to the size they have to split the program into overlays and thus have to use PLINK V 1.30. Unfortunately they get an error message regarding the object record type 8E What we like to know is: Is there a new version of PLINK available which supports MS-C 3.0 FULLY ? Specifically does it support the object record type 8EH fully? The German vendor expects them to pay about 300 $ without beeing able to tell them wether it does support MS-C or not. They told me the latest available version here in Germany was 1.48. Any help would be greatly appreciated. (Either to info-ibmpc or to IU%DACTH51.BITNET@WISCVM.ARPA or IU%DACTH51.BITNET@psuvax1.UUCP) el (Eberhard W. Lisse) ------------------------------ Date: Thu, 6 Mar 86 18:23:22 PST From: Matthew J Weinstein <matt@LOCUS.UCLA.EDU> To: info-ibmpc-request@isib Subject: MS C 3.0 Large model struct assign code generation BUG There are a number of known bugs in MS C 3.0 struct/double assignments. I found this one yesterday; given the following conditions: 1) you are generating large model code 2) you use ONE register variable in your function 3) you use that register variable to index a pointer to a struct 4) you are performing structure assignment of an element of that structure which is not the first elt of the structure If all of the above are true, the generated code contains two LEAs that are mis-ordered. Code that exercises the bug (try cc -S): ----- slash here ----- struct a1 {char foo[6];}; struct a2 {int bar; struct a1 elt;}; struct a2 *broken; struct a2 okay[]; Broken() { register i; struct a1 wrong; struct a2 *me_too = broken; wrong = broken[i].elt; /* broken */ wrong = me_too[i].elt; /* broken */ wrong = okay[i].elt; /* WORKS */ } ----- and here ----- No fix forthcoming from Microsoft. - Matt ------------------------------ Date: Fri 7 Mar 86 14:38:35-PST From: Carl Fussell <CARL%SCU%PANDA@SUMEX-AIM.ARPA> Subject: optimizer for DeSmet C compiler To: info-ibmpc@USC-ISIB.ARPA For those of you using the DeSmet C compiler... I've written a public-domain optimizer that is available from the C-Ware Fido Bulletin Board (143/9) at (408)-720-8197. The file is stored in the IBM PC file area under the name O88-V207.ARC (version 2.07 of O88). The ARC file contains three files: O88.DOC - The documentation file O88.EXE - O88 optimized by O88 for execution on an 8088 O88.V20 - O88 optimized by O88 for execution on a NEC V20. (rename O88.V20 to O88.EXE before use.) The optimizer automatically senses whether an 8088 or V20 is installed, and optimizes for the corresponding instruction set. Either mode can be forced by a command line argument. The V20 instructions that are used are all 80188 instructions, and so the optimizer is also useful for people with AT's. In general, it will remove 5-10% of the instructions (10-20% if the 80188 instruction set is used), and simplify ~5% of those that remain. One user said that the optimizer is better than that offered with MicroSoft's compiler. The optimizer is free for the taking. Enjoy! Dan Lewis Key Software Products ------------------------------ Date: Fri, 7 Mar 86 04:19:46 PST From: Ya'akov_Miles%UBC.MAILNET@MIT-MULTICS.ARPA To: info-ibmpc@USC-ISIB.ARPA Subject: `C` sources & MS-DOS implementation of MIT teco editor I am in the process of translating the MIT context editor TECO into standard Microsoft (UNIX) "C", and have a preliminary release of this editor available in both "C" source and (MS-DOS) executable form which is public domain, and can be downloaded at any time from the following 2400 baud RBBS bulletin boards shown (no user validation required) RBBS PoCo.............604-464-7693 RBBS Smokey Mountain..604-462-8753 These versions will be updated if I can continue with my project... ------------------------------ Date: Fri, 7 Mar 86 08:39:56 PST From: Ya'akov_Miles%UBC.MAILNET@MIT-MULTICS.ARPA To: info-ibmpc@USC-ISIB.ARPA Subject: Wish list for Microsoft C compiler It would be nice if Microsofts "C" compiler ver 3.0 was enhanced by (1) Improving the optimizer such that string constants are folded. Currently the following "C" statements PRINTF("\nThis string is stored twice by Microsoft `C`"); PRINTF("\nThis string is stored twice by Microsoft `C`"); causes the string constant to be replicated, wasting space. (2) Increasing the module size that the optimizer will handle before warning: function `xxxxxx` too large for post-optimizer [0] (3) Lowering the price sufficiently so as to become industry standard ------------------------------ Date: Fri, 7 Mar 86 09:18:10 est From: T. J. Brando <brando%linus@mitre-bedford.ARPA> To: INFO-IBMPC@USC-ISIB.ARPA Subject: RE: emacs-like editor I've been using an editor called BRIEF on an AT running PC-DOS 3.1 for about 4 months. I didn't realize until a couple of weeks ago that it's an emacs clone. When I did realize it I immediately switched from vi to emacs on the vax at work, and I'm in pig heaven! BRIEF uses ctrl-keys for movement commands and alt-keys for other commands. It supports multiple buffers, multiple files, and multiple windows. The key bindings of existing functions can be changed, and macros can be written in lisp-like format using most of the existing functions. In fact, the editor comes with a large number of user-contributed macros that were the winners in a contest the company sponsored. For more information on BRIEF, contact Solution Systems 335 Washington Street Norwell, MA 02061 Sorry I don't know how much it costs. Thom linus!brando@MITRE-BEDFORD.{ARPA,BITNET,CSNET} {allegra,decvax,ihnp4,philabs}!linus!brando.UUCP ------------------------------ Date: Fri, 07 Mar 86 13:01:58 EST From: ST602556%BROWNVM.BITNET@WISCVM.WISC.EDU To: Info-IBMPC@USC-ISIB.ARPA Subject: Emacs-like editor BRIEF, published by Solution Systems in Massachusetts, is a totally reconfigurable editor which is very much in the EMACS vein. Although it doesn't really start out too EMACS-like, you could very easily remap the keys and write a few macros so that it would be just like EMACS. I first used it last summer and within hours of getting it I was writing small macros. I don't use it heavily, but find it more than powerful and fast enough for the things I do. Eric Wolf ST602556@BROWNVM.BITNET P.S. - Although BRIEF's authors are friends of mine, I in no way profit from its sales. ------------------------------ From: op987!jsm@seismo.CSS.GOV To: info-ibmpc@usc-isib.arpa Subject: Queuing System Simulation Software Query Date: Thu Mar 6 08:16:51 1986 Has anybody seen a PC/XT or PC/AT package of simulation software that is useful for queuing-system simulation. Something like GPSS would be ideal, but anything close might wrok as well. The application involves modeling the queuing behavior of information management systems in which each node can be thought of as a single or multi-server queuing system. Can anybody help? Jim McMichael UUCP: seismo.css.gov!op987!jsm ------------------------------ Date: 7 Mar 1986 12:04:52 PST Subject: Re: FFT Boards, Frame-Grabber Hardware From: Billy <BRACKENRIDGE@USC-ISIB.ARPA> To: Ready.pa@XEROX.COM cc: info-ibmpc@USC-ISIB.ARPA We built a system based on the Imaging technology frame grabber and our TMS320 signal processor board. The board was built for speech processing, but we used the fast multiply to do cosine transform based image compression. Warner Frei sells this board along with a very nice mouse/window driven image processing system. He has FFTs running on the Marinco array processing board that will meet your needs. The Marinco board does floating point. I don't think 16 bit word is wide enough for 512x512 FFTs starting with 8 bit image data. Warner Frei (213)452-1730 ------------------------------ Date: 7 Mar 1986 14:15:13 CST Subject: FSCRC contract... From: HUNEYCUTT@GUNTER-ADAM.ARPA To: Info-IBMPC@USC-ISIB.ARPA In response to a general flood of info requests, this is a shotgun initial response. To those of you with no interest, I'm sorry to abuse your systems. The new contract (Z-248s) is not a GSA contract. Rather, its a requirements contract administered by the Air Force Computer Acquisition Center (AFCAC). AFCAC is copying the contract, and will be sending copies to all affected contracting offices. Check with you contracting office if in doubt. For those of you in the Air Force, your MAJCOM/SOA small computer technical center (SCTC) can assist you. By the way, two items I forgot to mention. First, the delegation of procurement authority has been modified to include all of the Department of Defense. Second, the systems come with a one (1) year warranty. Doug ------------------------------ Date: 7 Mar 86 13:27:48 PST (Friday) Subject: Incremental File Backup Utilities From: Burton.osbunorth@Xerox.COM To: info-ibmpc@usc-isib.Arpa A recent Digest article referred to IBU, an incremental backup utility, available for ftp from SIMTEL20. According to IBU's documentation, it determines files qualifying for backup by examining file dates and times on source and destination. If the source has a later date and time, then the file is marked for backup. (IBU also has options for overriding these selections.) Why not take the more straightforward of using the ARC bit? Consider the following: I have a system with a hard disk, and I organize my work by application area, each in a separate subdirectory. I also back up my work to floppy including full directory path. For some applications, I have only a few files, and I can back up on one floppy, or even share a floppy with another application. However, some applications have megabytes of files, and the backup spans multiple floppies. A date comparison method won't work automatically. Any pointers to backup utilities that use the ARC bit? Ideally, the utility is written in C and scans all subdirectories, not just the working directory. Phil Burton, Xerox Corp. ------------------------------ From: Tim Finin <Tim%upenn.csnet@CSNET-RELAY.ARPA> Subject: turbo prolog To: info-ibmpc@usc-isib.arpa Date: Fri, 7 Mar 86 08:42 EST Someone gave me a copy of a news item from Electronic Engineering Times of March 3rd which describes a Prolog compiler for PCs that Borland International (of Turbo Pascal fame) is releasing on April 15th. According to the note, the price will be $99. Borland claims that it was clocked at 100K lips on an IBM-PC and 300K lips on an AT! (The benchmark used was described as "a single rule benchmark"). The dialect is described as "a superset of Clocksin and Mellish". The system appears to include an incremental compiler, screen editor, support for windowing, a module capability, sound primitives and color graphics primitives. ------------------------------ Date: Sat, 8 Mar 86 11:39:11 est From: ucdavis!lll-crg!seismo!cmcl2!phri!greenber@ucbvax.berkeley.edu (Ross Greenberg) To: ucdavis!info-ibmpc Subject: Re: DOS Termination Regarding the DOS Termination problem from a program in C. The program as listed should correctly set the termination code.... and then, when you fall off the end of main, a new termination code is generated by main (_main?). Try simply putting: exit(code); at the end of the routine, or try this after setting your exit code: regset.x.ax = 0; intdos(®set, ®set); which will cause your code to exit immediately, without any of the stuff a the end of main being called. ross ross m. greenberg ihnp4!allegra!phri!sysdes!greenber [phri rarely makes a guest-account user a spokesperson. Especially not me.] ------------------------------ To: Gary Chapman <CHAPMAN@nyu-acf7.ARPA> cc: info-ibmpc@usc-isib.ARPA Subject: Re: Priam Drives on AT, Wait State on AT Date: Fri, 07 Mar 86 08:56:11 -0500 From: Dan Grim <grim@huey.udel.EDU> I have had the Priam 60MB drive installed in my AT for about two weeks. It works quite well with one reservation. The software supplied with the drive uses a non-standard partitioning scheme to take full advantage of the size of the drive. Although they configure the AT for a type 12 drive (which has 855 cylinders), the Priam drive actually has 981 cylinders. They use an installable disk driver under DOS to access the extra area. Since I am running both DOS and Xenix on my AT I am currently using the 60MB drive as a standard type 12 (50MB) drive. If you only use DOS this will cause you no problem. Dan ------------------------------ Date: Sat, 08 Mar 86 18:45:20 EST From: James H. Coombs <JAZBO%BROWNVM.BITNET@WISCVM.WISC.EDU> To: INFO-IBMPC@USC-ISIB.ARPA Subject: Limits on files open: concurrent and lifetime. A number of people have asked recently about the maximum number of files that a program may open. Since no one else has picked up the ball, I will offer what little I know about this. 1) The DOS manual states that no program can have more than 20 files open concurrently. This includes such files as stdio and stderr. Using the C fopen() function, I bomb out on the 16th file. 2) Some people have said that the limit is not 20 files CONCURRENTLY but 20 files during the execution of a program. Using standard C routines, I was able to open, write to, and close 40 files sequentially. 3) I tried a couple of editors. One could open only 15 files; the other opened 23 and appeared to be ready for many more. There were no indications that the second editor was paging files to disk. Clearly, there is some limit on the number of files that may be open concurrently, and no apparent (or documented) limit on the number that may be open during a complete run. Furthermore, there are techniques to get by the limit on the number of files open concurrently. Apparently, the 20 file limit came with DOS 2.0 and the use of file handles instead of control blocks. I tried using interrupt 21 function 0x0f to open files with fcbs but had no luck. I imagine that I was making some simple error, however, because this is the way fileio was handled before DOS 2.0. Conclusions: 1) An inability to open more than 20 files during the run of a program must have to do with some quirk/bug in either the compiler/interpreter that one is using or in DOS 3.xx. (I was using 2.0.) Will someone with DOS 3.xx check this out? 2) Programmers have been able to get around the 20 concurrent files limit, probably by using fcbs. This may be as simple as using the "archaic" libraries of compilers that support DOS 1.xx. If one does not have such a compiler, try using the DOS macros for fcbs. In the worst case, one could use direct reads and writes, but it is probably not necessary to give up that much of OS services. For more info, see the DOS manual, the Technical Reference, and the Programmer's Guide to MS-DOS. ------------------------------ Date: 21 Feb 86 20:58:00 PST From: Eliot Moore <Swg.Elmo@USC-ISIB.Arpa> To: info-ibmpc Subject: LSPELL added to Library LSPELL 0.93 LSPELL.PAS LSPELL is a TEST version of a simple but useful document LSPELL.DOC speller based on Borland International's Turbo Lightning. Christopher J. Dunford The Cove Software Group 10057-2 Windstream Drive Columbia, MD 21044 (301) 992-9371 CompuServe 76703,2002 12/10/85 ------------------------------ Date: 8 Mar 1986 23:40:31 PST Subject: Dirlib Package added to Library From: Eliot Moore <SWG.ELMO@USC-ISIB.ARPA> To: info-IBMPC@USC-ISIB.ARPA dirlib.intro Dirlib is the "dirlib" package implemented for dirlib.c Microsoft-C. Matt@Ucla-Locus, 25-Feb-86. dirlib.h testdir.c ------------------------------ Date: 9 Mar 86 06:58:33 EST From: Davidann <DZimmerman@BLUE.RUTGERS.EDU> Subject: Kermit 2.28 jrd/5b + Turbo Lightning To: Info-IBMPC@USC-ISIB.ARPA I've noticed problems also with the combination of Kermit and Lightning. Normally, Lightning is pretty well behaved, and even with the smallest dictionary in memory (RAM1), it doesn't beep unless a word looks adequately strange. Since I started running it with jrd/5b, though, it has been beeping on little words like "the" and "my", which is totally out of character. If I then do a word check, Lightning tells me the word is correct without accessing the disk, which means #1: the word was already in memory, and #2: if it was in memory, Lightning shouldn't have beeped. I should mention that Lightning was running fine with my previous version of Kermit (2.28 jrd/2). David P. Zimmerman, Rutgers University Arpa: DZimmerman@Blue Uucp: {seismo, allegra, ihnp4}!topaz!dpz ------------------------------ Date: Fri, 7-Mar-86 10:37:28 PST From: ihnp4!pelican!pete@ucbvax.berkeley.edu (Pete Carah) Subject: /G0 and /G1 benchmarks To: Fishwick%upenn.csnet@csnet-relay.arpa Cc: info-ibmpc@usc-isib.arpa >Has anyone done any benchmarks on the code generated from using the >options /G0, /G1, or /G2 under Microsoft's C V3.0 ? I have run (on a NEC V20, not a 186) the Dhrystone benchmark using both the G0 and G1 options, with and without /Oxxx, and with and without the REGISTER as given in the dhrystone header. G1 does speed things up. Note that this benchmark is very call-intensive, simulating an ADA environment with highly-structured, small routines. Here is my letter to the dhrystone keeper giving the results. I also ran it on the definicon board (rather impressive). Note that the times differring by 1 second are not particularly significant due to the time() granularity of 1 second. The number listed alone or before seconds is "dhrystones/sec" (= 50000/seconds). I include some of the other times that came with the benchmark for reference: no reg reg * IBM PC/XT 8088-4.77Mhz VENIX/86 2.0 cc 297 324 * VAX 11/750 - VMS VAX-11 C 2.0 958 1091 * Sun2/120 68010-10Mhz Sun 4.2BSD cc 1136 1219 * PDP 11/70 - UNIX 5.2 cc 1162 1250 * IBM PC/AT 80286-7.5Mhz VENIX/86 2.1 cc 1190 1315 * * VAX 11/780 - UNIX 5.2 cc 1515 1562 I ran dhrystone on my PC in various combinations of compile options, and got the following results: PC (not XT) 4.77 MHz NEC V20 (not 8088), MSDOS 3.1, Microsoft C v.3.00 This compiler does support enums and structure assigns. I used small model for all of these (2-byte pointers). no regs, no optimize, no 186 instructions: 387 129 secs no regs, no optimize, 186 instructions: 393 127 secs no regs, -Ot, no 186 instructions: 387 no regs, -Ot, 186 instructions: 393 regs, no optimize, no 186 instructions: 420 119 secs regs, no optimize, 186 instructions: 427 117 secs regs, -Ot, no 186 instructions: 416 120 secs regs, -Ot, 186 instructions: 427 117 secs Apparently the -Ot option did nothing and enabling the 80186 instruction set improved things by a couple of percent. Same PC, DSI-32 board, 10 Mhz 32032, using the Definicon (Green Hills) C: This compiler supports enums and structure assigns. The base system was MSDOS 3.1 The runtime version was loader: 2.14v; 32io: 2.1 No wait states: Definicon 32032, no regs, no optimize: 1282 39 secs Definicon 32032, no regs, optimize 1: 1250 40 secs Definicon 32032, regs, no optimize: 1315 38 secs Definicon 32032, regs, optimize 1: 1282 39 secs 1 wait state: (my 2MB upgrade was slightly flaky with none) Definicon 32032, no regs, no optimize 1190 42 secs Definicon 32032, no regs, optimize 1190 42 secs Definicon 32032, regs, no optimize 1250 40 secs Definicon 32032, regs, optimize 1190 42 secs (!) A different PC, 8088, 4.77 MHz PC-IX, cc: (cron and update were left running, with no crontab activity) no regs 271 regs 294 -- Pete ...{ihnp4, scgvaxd, vortex}!pelican!pete ------------------------------ Date: 3 Mar 86 13:43 EST From: Jim Moore <moore@ncsc> Subject: NANSI Query To: info-ibmpc@usc-isib Is NANSI available anywhere as executable rather than source? I don't have Microsoft's MASM 3.0, and apparently that's what I need for the current stuff available in ISIB (NANSI.LBR). Thanks for any information jim [NANSI is the new ANSI.SYS replacement from Daniel Kegel of Cal Tech -wab] ------------------------------ Date: 5 Mar 1986 13:52:33 PST Subject: NANSI From: Billy <BRACKENRIDGE@USC-ISIB.ARPA> To: Jim Moore <moore@NCSC.ARPA> cc: info-ibmpc@USC-ISIB.ARPA There were some modifications to the code receintly that allow it to assemble under MASM 1.00. We won't support binary, but somebody else on the net might. ------------------------------ Date: Tue 4 Mar 86 14:06:21-PST From: Marvin Zauderer <ZAUDERER@SU-SUSHI.ARPA> Subject: Turbo's 64K Limitation Query To: info-ibmpc@USC-ISIB.ARPA I'm interested in getting around Turbo Pascal's 64K code/64K data limitations. I know about the "chain" and "execute" calls; does anyone have any experience with other methods, e.g. commercial products which extend Turbo? Thanks for your help. -- Marvin Zauderer E-Mail: Zauderer@SU-SUSHI.ARPA ------------------------------ Date: Wed, 5 Mar 86 23:29:31 pst From: Marc Caminetsky <marc%uofm-uts.cdn%ubc.csnet@CSNET-RELAY.ARPA> To: info-ibmpc@usc-isib.ARPA Subject: Compaq Deskpro and NEC V30 Query I have just purchased a COMPAQ deskpro and was thinking about poping in the NEC V30 chip to replace my 8086. I have many friends who have replace there 8088 with a V20 in their IBMPC and have met with a vast improvment in speed. Firstly, can you put the V30 in the DESKPRO and secondly has anyone done it? If so what kind of speed improvments is it giving? Thanks in advance, Marc Caminetsky. ------------------------------ Date: Wed, 5 Mar 86 23:29:42 pst From: Marc Caminetsky <marc%uofm-uts.cdn%ubc.csnet@CSNET-RELAY.ARPA> To: info-ibmpc@usc-isib.ARPA Subject: Qubie Modem Problem I have a Qubie 1200 baud internal for my Compaq Deskpro, It works fine Usually, in fact I am using it now. BUT..... I have tried to logon to a certain BBS that uses a multitek 2400 baud modem. When I dial, I connect, hear the host modem at 2400 baud, it then switches to 1200 baud and then nothing happens. The QUBIE just sits there and does nothing. It will not respond to the Multitek carrier. Has anyone else had this problem with there qubie's? Is there some way via the AT commands that I can rectify this problem? Other than this particular BBS the modem works fine....I should also mention that a friend uses a HAYES 1200 and has NO prob. Thanks in advance Marc Caminetsky. ------------------------------ Date: 6 Mar 86 09:10 GMT From: oli2146 @ KOREA-EMH Subject: Lotus 1-2-3 Ver 1 vs. Ver 2 Query To: info-ibmpc @ usc-isib.arpa Can anyone send me or direct me to a comparison of Lotus 1-2-3 Version 1 and version 2? Thanks in advance.... Mark Meaders "Fm the Land of the Morning Calm...." ------------------------------ Date: Thu, 6 Mar 86 12:23 EST From: Mandell@DOCKMASTER.ARPA Subject: GEM Draw Problem To: info-ibmpc@USC-ISI.ARPA I saw a demo of GEM draw last week. The use of the shift home combination to select the movement of objects(open hand) seemed to be very unreliable. It requires a simultaneous momentary push othe two keys and only works a small portion of the time. Has anyone seen this problem? is there a fix? Does a similar problem occur with a mouse? Dick Mandell ------------------------------ Date: Thursday, 6 Mar 1986 09:04:00 PST From: (Dave Gomberg) <GOMBERG@UCSFVM.BITNET> To: info-ibmpc at USC-ISIB Subject: ICOM R71A Radio Receiver to PC Query Has anyone succeeded in connecting an IBM PC to the "computer" interface of this beast? [This is a ham radio receiver with a serial interface that allows selection of frequency etc. -wab] ------------------------------ Date: Thu, 6 Mar 86 15:41 EST From: TRK%OAK.SAINET.MFENET@LLL-MFE.ARPA Subject: CMS Hard Disk, IBM Proprinter, Write & Spell Query To: Info-IBMPC@USC-ISIB.Arpa My father is in the market for an XT system for his club for keeping track of names & addresses, sending out letters etc. He is not a computer person and neither are the people who will be using it. What he is looking at is an XT with a 20 meg CMS hard disk, an IBM proprinter, and a Write & Spell word processor. I have never used anything by CMS and I am not familliar with the printer or the word processor either. I was hoping that some of you out there might be able to tell me if you have had any good/bad experiences with these things so I can tell my father whether to go ahead and get it or keep looking for something better. I would certainly appreciate any help anyone could offer. You can send replies to Info-IBMPC or: TRK%OAK.SAInet.MFEnet@LLL-MFE.Arpa Thanks in advance, --Tracy ------------------------------ Date: Thu, 6 Mar 86 16:15:25 CST From: frank@db.wisc.edu (Dan Frank) To: info-ibmpc@usc-isib Subject: Ctags for MS-DOS wanted Does anyone know of a public domain ctags program for MS-DOS? ------------------------------ Date: Fri, 7 Mar 86 01:32:07 EST From: Jacob Moskowitz <JMSK@MC.LCS.MIT.EDU> Subject: Character ROMS Query To: Info-ibmpc@USC-ISIB.ARPA I am very urgently looking for alternate ROM's to plug in place of the one on a PC to give me more greek characters and more math symbols. A companion ROM for the printer would be nice, but I can download those. I've been told that there is an EPROM equivalent for the PC character ROM, but I have been unable to find a vendor that sells anything like what I want. The reason I can't use any of the numerous "scientific/math" WP pgm's is that I need to use these fonts in my existing sfw, ThinkTank in particular, and I sure don't need those foreign characters nor the smiling faces, etc. Any help any of you can give me on this would be greatly appreciated. Jake Moskowitz CBS, NYU Physics ------------------------------ Date: Fri, 7 Mar 86 12:42:38 pst From: calma!thompson@ucbvax.berkeley.edu (Sheldon Thompson) To: ucbvax!info-ibmpc Subject: Serial printer on PCjr Query I desparately need help! I am attempting to access the serial port on the IBM PCjr to use a serial printer. The tech manual seems to be unavailable in this area. Any help appreciated. Thanks in advance, Sheldon Thompson Calma Company UUCP: ...{ucbvax,sun,csd-gould}!calma!thompson ARPA: calma!thompson!ucbvax.ARPA Disclaimer: I'm sure much of the above is trademarked. ------------------------------ Date: Fri, 7 Mar 86 13:51:46 PST From: atm%DEImos.Caltech.Edu@Hamlet.Caltech.Edu Subject: Real-time PC operating systems query Has anyone any experience with the various real-time operating systems for the PC, either full-featured ones such as RTOS (derived from iRMX86) or bare kernels such as VRTX? In particular, has anyone spliced an Ethernet protocol server (e.g. TCP/IP) into one of these? Alan Moffet Caltech Astronomy ATM%deimos@cit-hamlet.arpa ATM%deimos@caltech.bitnet ------------------------------ Date: Fri 7 Mar 86 14:53:07-PST From: STEINBERGER@SRI-KL.ARPA Subject: serial ports query To: INFO-IBMPC@USC-ISIB.ARPA I am using IBM professional fortran and need to control 2 serial ports. I have ordered a product called "No Limit" from MEF ENVIRONMENTAL in Austin, TX (800) 562- 9700. They claim that their software will allow access to both ports, including using interrupt driven communication. Has anyone used this or a similar product? Any advice or comments would be welcome. Thanks. Ric Steinberger STEINBERGER@SRI-KL (415) 859-4300 [There are a couple of programs in the INFO-IBMPC library that do this. -ed] ------------------------------ Date: Sat 8 Mar 86 13:36:10-PST From: Carl Fussell <CARL%SCU%PANDA@SUMEX-AIM.ARPA> Subject: graphics query To: info-ibmpc@USC-ISIB.ARPA Does anyone have an assembler routine that can be interfaced with the Professional FORTRAN to run the EGA board on the IBM AT using a monochrome monitor??? I've written some code that allows me to turn pixons on and off, but I was also looking for some assembler routines that will allow me to display characters and graphics (Mode F) simultaneously. I've also been hacking at some code to allow the user to define an alternate character set, but have had some trouble interpretting the limited documentation. If anyone has any assembler routines that do this and can be interfaced with the Professional FORTRAN package, I'd like to hear from you. Thanks, Stephen J. Hager (hager%SCU%Panda@Sumex-Aim.Arpa) ------------------------------ Date: Sat, 08 Mar 86 18:53:59 EST From: James H. Coombs <JAZBO%BROWNVM.BITNET@WISCVM.WISC.EDU> To: INFO-IBMPC@USC-ISIB.ARPA Subject: Clearing the screen query What is the best way to clear the screen when running the color monitor? The call in my C library wipes out the attribute bytes. I believe that this is a call to the BIOS video services. As I understand it, one can print an ANSI.SYS command string to clear the screen, but this requires the installation of ANSI.SYS. If possible, I would like to make no assumptions about the environment. There must be some generic way to clear the screen (better than printing 25 blank lines). ------------------------------ Date: Sat, 8 Mar 86 19:50:13 EST From: Jacob Moskowitz <JMSK@MC.LCS.MIT.EDU> Subject: Alternate Character Sets Query To: info-ibmpc@USC-ISI.ARPA I just posted a message requesting help on replacing the standard character ROM, in particular with one adding to the limited number of greek/math symbols. Since then, a friend told me that STSC sells an APL compiler that changes some of the characters to APL characters, WITHOUT replacing the ROM, suggesting that there's a way to do this by software. Does anyone know whether this is true ? Does the IBM-PC download the character ROM into RAM, or does DOS have to be patched to allow intercepting of interrupts, or something ? I'm obviously ignorant of this stuff and will be very grateful for any assistance anyone out there can offer. Thanks. Jake Moskowitz ------------------------------ Date: 9 Mar 86 21:10:52 +0100 From: XBR1YD14%DDATHD21.BITNET@WISCVM.WISC.EDU (YD14@BR1.THDNET) Subject: Program chaining query To: info-ibmpc@usc-isib I have some FORTRAN programs that should invoke each other, so that only one program is resident in memory at any time. There is a DOS-call in version 3.x of DOS, but I don't like it because it seems to create some kind of subprocess and the "calling" program stays in memory. (In the CP/M-80 version of my program I'm using the FCHAIN subroutine in the F80's library.) I'd appreciate any hints. If ever possible, I'd like to use DOS Version 2.11. Thank you Reinhard Goeth (Techn. Univ. of Darmstadt) BITNET: XBR1YD14@DDATHD21 ARPA: XBR1YD14%DDATHD21.BITNET@WISCVM.WISC.EDU ------------------------------ Date: 9 Mar 1986 15:59-EST Subject: Wang PC "Open Architecture" package query From: ABN.ISCAMS@USC-ISID.ARPA To: info-ibmpc@USC-ISIB.ARPA Has anyone seen or worked with the "Open Architecture" package (or some such name) obtainable (for much bucks) from Wang for their PC? It supposedly has technical information about the buss, BIOS, interrupts, and all the secret ways it is NOT compatible with most anything. They blindsided us when we bought the boxes .. I'd rather not have any more unpleasant surprises if I buy this supposed equivalent to "Norton's Programmers Guide" for the Wang PCs. Thanks in advance, David Kirschbaum Toad Hall ABN.ISCAMS@USC-ISID.ARPA ------------------------------ End of Info-IBMPC Digest ************************ -------