shawn@bmcg.UUCP (Shawn Bosley) (05/18/86)
In article <540@gould9.UUCP> joel@gould9.UUCP (Joel West) writes: >I'm still collating info for UNIX on a xxx86 machine (and would note >that I got a (postal) mailing from IBM in Austin about Xenix 2.0, based >solely on a previous posting.) > >I've narrowed myself to a 286-based system, probably a $3-$4k clone. >I definitely will be buying something that is still bundled, true >to the UNIX ideal. I have two quick questions: > >1. Which UNIX or XENIX take advantage of the protected mode? (presumably > to prevent a misbehaving task from crashing the whole cpu). IBM's XENIX (version 2.0) is derived from UNIX SYSTEM V and has everything you need to work well in UNIX. Version 1.0 had some horrible problems with it (I had to suffer through most of them) but version 2.0 seems to have cleared up the limitations (they should have never released version 1.0!) XENIX runs is the protected mode so that it would be very difficult to crash the system. I would recommend XENIX because it has good support and I haven't had too many problems porting code from other UNIX systems over to it. > >2. Does the 286 chip still have that awful segmentation small model > vs. large model crap? Or, are there 286 instructions which make > it possible to address > 64kb contiguously? If so, does the > C compiler for #1 support this? > Unfortunately, I doubt that you could elimnate this problem because of the inherent design of the 286 chip. However, XENIX offers small, medium, and large models for C programs. It's really not too inconvenient in the sense that you *know* what model is necessary to do the job. >Any 286 *NIX benchmarks against a VAX would also be interesting... >-- > Joel West (619) 457-9681 > CACI, Inc. Federal, 3344 N. Torrey Pines Ct., La Jolla, CA 92037 > {cbosgd, ihnp4, sdcsvax, ucla-cs} !gould9!joel > joel%gould9.uucp@NOSC.ARPA --Shawn Bosley
davidsen@steinmetz.UUCP (Davidsen) (06/10/86)
In article <2082@bmcg.UUCP> shawn@bmcg.UUCP (Shawn Bosley) writes: >In article <540@gould9.UUCP> joel@gould9.UUCP (Joel West) writes: >>I'm still collating info for UNIX on a xxx86 machine (and would note >>that I got a (postal) mailing from IBM in Austin about Xenix 2.0, based >>solely on a previous posting.) >> >>I've narrowed myself to a 286-based system, probably a $3-$4k clone. >>I definitely will be buying something that is still bundled, true >>to the UNIX ideal. I have two quick questions: >> >>1. Which UNIX or XENIX take advantage of the protected mode? (presumably >> to prevent a misbehaving task from crashing the whole cpu). Both IBM and SCO flavors of XENIX use protected mode. I believe that the UNIX from Microport does also. >>2. Does the 286 chip still have that awful segmentation small model >> vs. large model crap? Or, are there 286 instructions which make >> it possible to address > 64kb contiguously? If so, does the >> C compiler for #1 support this? >> Fact of life with the 286. SCO XENIX C has a "huge model" which uses long subscripts and treats memory as a flat addressing space. It has worked for me on several projects, but I am *not* claiming it is bug free. SCO XENIX supports ten virtual terminals so you can be doing communications on one, switch to another to read mail, another to look up a phone number of use a database, etc. The IBM version doesn't seem to have this. Comments on XENIX: the C compiler is *NOT* based on PCC. It has some reasonable optimizing, generates code for many models, and has far better error checking than PCC, but it won't compile some UNIX programs. Suggestion: Microport UNIX for $350, having growing pains but uses real PCC and uucp. This is a somewhat of taking a chance, but all of the parts are SysV, rather than custom. Why not buy a cheap 68k based system? Megadata is advertizing the hell out of their $3200 system, and an ATT7300 with UNIX is about $5K (less 25% typically). I added a 32MB Tulin ($650) to mine and it runs great. Lotsa good stuff around. Enjoy! -- -bill davidsen ihnp4!seismo!rochester!steinmetz!--\ \ unirot ------------->---> crdos1!davidsen / sixhub ---------------------/ (davidsen@ge-crd.ARPA) "Stupidity, like virtue, is its own reward"