plocher@uport.UUCP (John Plocher) (10/05/88)
( Marketing Hype follows; This is the only "marketing" I will do here for these products. I am posting a heavily edited version of the press release, call Microport Sales (800) 722-8649 to obtain a complete upgrade anouncement. Sorry if this offends anyone, my email address is in the header -John ) Contact David Le, Microport Systems, (408) 438-8649 Microport unveils new Unix (r) System V releases for 80286 and 80386 micros Scotts Valley, Calif. -- September 6, 1988 -- Microport Systems, Inc announced today that it will start shipping two new releases of its Unix System V products during the month of September: System V/386 version 3.0e for Intel 80386-based computers and System V/AT version 2.4 for 80286-based machines. An improved version of Dos Merge 386, a program which runs DOS applications under 80386 Unix systems, was also announced. ... The new features provided in System V/386 version 3.0e include: Support for RLL and ESDI hard disk controllers, support for streaming tape backups and for the Weitek floating point coprocessor; a 32 bit, feature rich, highly optimizing Greenhills C compiler; an improved, menu driven user interface called SysViz, added utilities; and improved documentation. ... Existing users of Microport's System V/386 can upgrade to the new 3.0e version by purchasing new Runtime disks and release notes for $99, a new Runtime Manual for $50, and a new Software Development System disk set for $125. A new release of Dos Merge 386, version 1.1, was also announced. ... Current users of Dos Merge 386 will obtain a free upgrade to version 1.1 with the purchase of the V/386 Runtime disk pack upgrade. System V/AT ... will be released as version 2.4. ... System V/AT version 2.4 offers many new improvements including: support for RLL & ESDI hard disk controllers and streaming tape backups; improved mail program; improved floppy and fixed disk drivers, more utilities and new vi functions; and new software development libraries. ( HDB uucp is also included -John ) ... Upgrades are avaliable to existing users at $75 for new Runtime disks and Release notes, and $99 for the Software Development System disk set. Microport Systems 10 Victor Square Scotts Valley, CA 95066 (408) 438-8649 (r) Unix is a registered trademark of AT&T
nusip@maccs.McMaster.CA (Mike Borza) (10/06/88)
This is all very nice, but both your blurb, and the promo sent out to registered users fails to say whether SV/AT 2.4 includes the '287 exception hang fix that's in my Beta 2.3 kernel-- the fix for system hangs on FP exceptions when a '287 is installed. What's the story? mike borza <nusip@maccs.uucp>
scotty@l5comp.UUCP (Scott Turner) (10/09/88)
In article <12950@mcdchg.UUCP> plocher@uport.UUCP (John Plocher) writes: >( Marketing Hype follows; Carefully edited marketing hype, John edited out the part about Microport not living up to the outstanding upgrade contracts. This site at least paid $$$ for automatic upgrades on up to 2 major releases in a one year period and 10 free bug fixes, all postage paid. Instead we now find that Microport demands: 1. PROOF OF OWNERSHIP!!! We must return diskettes to Microport to prove we own it, I thought we did that when we bought the damn update contract??? 2. We have to pay them $15 postage. We already paid $$$ for postage, now they want to stiff us for another $15??? 3. The "automatic" upgrade just isn't too damn automatic anymore... Item 2 is so petty I just can't understand what Microport is up to. We've already paid Microport MORE than enough to cover the physical costs of the upgrade. Also, there is no mention that I can find about those poor souls who just recently (within last 60 days lets say) purchased System V/386 2.2. Do they get free upgrades? Do they get reduced price upgrades? Or do they get to pay full price for an upgrade? For a company that has otherwise treated at least this customer quite well, I can't understand this shift (or should I say shaft? :) in position. Also, as regards the Greenhill's compiler, we haven't seen any major speed changes using this compiler. The compress program for example was actually slower compiled with gcc than with the standard cc under certain conditions. (and we were never able to get gcc to make a compress much faster than one from cc -O) Also, those expecting to use the Greenhill's compiler on GNU software may be in for an unpleasant shock. The gcc compiler does NOT reload registers at the end of a routine based on an offset off the frame pointer, it just pops 'em off the stack. Which isn't a problem unless you are using alloca()... (something GNU software does ALOT of it seems) [NOTE: GNU and Greenhill's both call their C compilers gcc, when I say gcc in this article I mean the Greenhill's product] So far Microport hasn't supplied any work around for this little detail (surely this is a command line switch???) The Greenhill's compiler is roughly about twice as fast as the standard cc though. But so far this is the only speed advantage we've seen. Scott Turner scotty@l5comp -or- uunet!l5comp!scotty
zeeff@b-tech.ann-arbor.mi.us (Jon Zeeff) (10/11/88)
In article <444@l5comp.UUCP> scotty@l5comp.UUCP (Scott Turner) writes: > >Also, as regards the Greenhill's compiler, we haven't seen any major speed >changes using this compiler. The compress program for example was actually >slower compiled with gcc than with the standard cc under certain conditions. >(and we were never able to get gcc to make a compress much faster than one >from cc -O) > I find Green Hill's gcc to compile much quicker and produce code about 20% faster (even on compress). True, it doesn't work with some programs - GNU stuff, news 2.11, ksh. Now if I can only get GNU gcc 1.29 to work on the '386. -- Jon Zeeff Branch Technology, umix!b-tech!zeeff zeeff@b-tech.ann-arbor.mi.us
james@bigtex.uucp (James Van Artsdalen) (10/12/88)
In article <4811@b-tech.ann-arbor.mi.us>, zeeff@b-tech.UUCP (Jon Zeeff) wrote: > I find Green Hill's gcc to compile much quicker and produce code about > 20% faster (even on compress). True, it doesn't work with some > programs - GNU stuff, news 2.11, ksh. Now if I can only get GNU gcc > 1.29 to work on the '386. This seems to raise the interesting question of what Green Hill's C *does* work with! As for GNU C, I have v1.29 up and running from a first stage compile. Seems to work compiling sendmail 5.59 anyway. Please post a comparison when possible. I don't think GNU C compiles any quicker than PCC (with the optimizer anyway), but it produces vastly better code. -- James R. Van Artsdalen ...!uunet!utastro!bigtex!james "Live Free or Die" Home: 512-346-2444 Work: 338-8789 9505 Arboretum Blvd Austin TX 78759