naim@nucsrl.UUCP (Naim Abdullah) (01/04/88)
A friend of mine has a Compaq/386 and is looking around for a
XENIX system that would allow him to access the largest amount
of virtual memory. His primary application is running fortran
programs that need huge amounts of memory ( >10Mb).
Could anybody recommend a XENIX system for his needs ? Does anybody
know how much virtual memory SCO Xenix for the Compaq 386 allows you to
access ? Does their fortran compiler allow you access that memory (or
can you only get to it via C) ? No reviews that I have seen address this
point and my friend would like to know before he sinks his money into
the OS.
Naim Abdullah
Dept. of EECS
Northwestern University
Uucp: {ihnp4, chinet, gargoyle}!nucsrl!naim
Internet: naim@eecs.nwu.edudavidsen@steinmetz.steinmetz.UUCP (William E. Davidsen Jr) (01/06/88)
In article <3700001@nucsrl.UUCP> naim@nucsrl.UUCP (Naim Abdullah) writes: | A friend of mine has a Compaq/386 and is looking around for a | XENIX system that would allow him to access the largest amount | of virtual memory. His primary application is running fortran | programs that need huge amounts of memory ( >10Mb). That's not much these days. | | Could anybody recommend a XENIX system for his needs ? Does anybody | know how much virtual memory SCO Xenix for the Compaq 386 allows you to | access ? Does their fortran compiler allow you access that memory (or | can you only get to it via C) ? No reviews that I have seen address this | point and my friend would like to know before he sinks his money into | the OS. In the order you asked: it sounds as if he has already bought a system, now all he needs is Xenix/386 and more memory. Virtual address space is a function of the swap size and physical memory. A high ratio of logical to physical memory in a program will usually result in poor performance. The memory is managed by the o/s, language doesn't matter. -- bill davidsen (wedu@ge-crd.arpa) {uunet | philabs | seismo}!steinmetz!crdos1!davidsen "Stupidity, like virtue, is its own reward" -me
dyer@spdcc.COM (Steve Dyer) (01/06/88)
In article <8364@steinmetz.steinmetz.UUCP>, davidsen@steinmetz.steinmetz.UUCP (William E. Davidsen Jr) writes: > In article <3700001@nucsrl.UUCP> naim@nucsrl.UUCP (Naim Abdullah) writes: > | Could anybody recommend a XENIX system for his needs ? Does anybody > | know how much virtual memory SCO Xenix for the Compaq 386 allows you to > | access ? Does their fortran compiler allow you access that memory (or > | can you only get to it via C) ? > The memory is managed by the o/s, language doesn't matter. Careful there, Bill. Although you're "right", you might be a little misleading. I'm not sure that there is a Fortran compiler available yet for XENIX 386 which generates 386 code. If our user isn't careful, he might go out and purchase a version of Microsoft Fortran for XENIX 286 (which would run quite nicely on XENIX 386, I'd wager) but I am not at all sure that product would give him seamless access to large data areas. Remember, the UNIX "f77" compiler doesn't come with XENIX 286 or 386. Maybe a Microsoft or SCO person could clarify this. It might be that Microport or ISC's offering, if either of them provides an F77 compiler for the 386, would win out, provided they can also satisfy the virtual memory requirements, and I would be surprised if they didn't. -- Steve Dyer dyer@harvard.harvard.edu dyer@spdcc.COM aka {ihnp4,harvard,husc6,linus,ima,bbn,m2c}!spdcc!dyer
davidsen@steinmetz.steinmetz.UUCP (William E. Davidsen Jr) (01/12/88)
In article <536@spdcc.COM> dyer@spdcc.COM (Steve Dyer) writes: | > | Could anybody recommend a XENIX system for his needs ? Does anybody | > | know how much virtual memory SCO Xenix for the Compaq 386 allows you to | > | access ? Does their fortran compiler allow you access that memory (or | > | can you only get to it via C) ? | Careful there, Bill. Although you're "right", you might be a little | misleading. I'm not sure that there is a Fortran compiler available yet | for XENIX 386 which generates 386 code. If our user isn't careful, | he might go out and purchase a version of Microsoft Fortran for XENIX 286 | (which would run quite nicely on XENIX 386, I'd wager) but I am not at all | sure that product would give him seamless access to large data areas. | Remember, the UNIX "f77" compiler doesn't come with XENIX 286 or 386. Thanks for clarifying this. My experience is that VM works for 286 programs, although they are in a segmented mode, limited to a "mere" 15MB, and run slower than the 386 version. Someone with Ryan-McFarland FORTRAN says that they allow access to a full 15MB address space. Corrections to the net if needed. I would love to have an f77 compiler, but wouldn't pay more than $100 for it, since I never write in fortran, and the only programs I have are games. -- bill davidsen (wedu@ge-crd.arpa) {uunet | philabs | seismo}!steinmetz!crdos1!davidsen "Stupidity, like virtue, is its own reward" -me