karl@ddsw1.UUCP (11/19/87)
This has probably been hashed out a million times before, but...
I'd like to ask the users of the net which 80386 Unix (Xenix) derivitive
they would buy. I know of the following companies producing a Unix(Xenix)
V/386 -- feel free to add to this!
o Microport (System V/386)
o Interactive (Same)
o SCO (Xenix V/386)
o Bell Tech (?)
We're looking for the following:
o Speed -- Performance is very, very important.
We will be developing code on this system -- thus, the quality of
the compiler(s) also count in a big way. Comments on both the system's
speed as a whole and on the quality of the code generated by the
compiler(s) is desired.
o Adherance to standards -- SVID is foremost, although utility-level
compatibility would also be nice.
o Reliability -- We cannot tolerate a product that panics or crashes.
We *need* 9600 baud communications ability, 19,200 is desired.
o Must handle 'smart' serial boards (requirement above mandates it).
o Must have a full link kit (ability to handle custom drivers *AND* tune
some kernel parameters); a section in the manuals that is worthwhile on
writing drivers for the particular varient of Unix is even more helpful.
o MUST (repeat, MUST) handle disk drives which are not in the standard
15-type set from an IBM AT. Soft configuration (ala Microport SysV/286)
is perfect, but any other configurable solution is also fine (I'm willing
to link a new kernel to change it, for example). What is *NOT* fine is
a scheme which limits you to those types which are in your BIOS. I believe
that this disqualifies Interactive (from their literature), but perhaps
that's not the whole story...
In addition, ability to run DOS/MERGE in some form is also desired, but not
necessary (we have other systems for that if needed).
We are primarially interested in comments from people who actually use the
product on an 80386 system. Advertising drivel and the like will go into
the trashcan -- but subjective comments ("Product 'x' feels slower than the
same company's '286 release") are welcome *as long as you have some basis in
reality*. *PLEASE* include specifics on the hardware and software you are
using so we can make some reasonably intelligent choices based on your
input. (ie: memory size is important, as is disk type/make/size, etc)
Pricing information is also appreciated, if you have it handy.... The more
complete the better!
Of course, we'll summarize responses to the net.
--
Karl Denninger UUCP : ...ihnp4!ddsw1!karl
Macro Computer Solutions Dial : +1 (312) 566-8909 (300-1200)
"Quality solutions at a fair price" Voice: +1 (312) 566-8910 (24 hrs)bill@ism780c.UUCP (11/23/87)
In article <361@ddsw1.UUCP>, karl@ddsw1.UUCP (Karl Denninger) writes: > This has probably been hashed out a million times before, but... > > I'd like to ask the users of the net which 80386 Unix (Xenix) derivitive > they would buy. I know of the following companies producing a Unix(Xenix) > V/386 -- feel free to add to this! > > ..... > o Speed -- Performance is very, very important. > We will be developing code on this system -- thus, the quality of > the compiler(s) also count in a big way. Comments on both the system's > speed as a whole and on the quality of the code generated by the > compiler(s) is desired. > When INTERACTIVE did the V.3/386 port, neither Intel nor AT&T were willing to fund a sophisticated globally optimizing compiler as part of the port. The RCC compiler does a very decent job of optimization, in some cases noticeably better than several well-known optimizing C compilers. It doesn't do as well on the Dhrystone. If performance is "very, very important", you should evaluate several compilers in addition to the RCC compiler to see which performs better for you and your application. Different customers have different requirements for their C compilers. The RCC compiler is likely the most reliable C compiler available for the 386 (we have tried 386 C compilers that would not correctly compile Unix commands or even worse, would apparently compile correctly but then produce incorrect results). Once again, it depends on your needs. For example, INTERACTIVE works with their OEM's to provide alternative C compilers, or other compilers, if the OEM has the requirement for a compiler in addition to the RCC compiler. Any C compiler that claims to work with "Unix 386" (as opposed to Xenix) should work fine with any of the available "Unix 386" systems. > o Adherance to standards -- SVID is foremost, although utility-level > compatibility would also be nice. > Once again, any of the available "Unix 386" systems should be SVID compatible. They are all based on the certified port that INTERACTIVE did for AT&T and Intel. If you have any doubt, ask your supplier whether their product is based on the "certified" Unix V.3/386 release, as opposed to an earlier beta release. All AT&T licenses allowing distribution of Unix V.3/386 mandates that a vendor must ship a product based on the certified release. The certified port has been available since July. Any vendor that is not shipping a product based on the certified V.3/386 port is going to have difficulty with AT&T. > o Reliability -- We cannot tolerate a product that panics or crashes. > We *need* 9600 baud communications ability, 19,200 is desired. > > o Must handle 'smart' serial boards (requirement above mandates it). > > o Must have a full link kit (ability to handle custom drivers *AND* tune > some kernel parameters); a section in the manuals that is worthwhile on > writing drivers for the particular varient of Unix is even more helpful. > > o MUST (repeat, MUST) handle disk drives which are not in the standard > 15-type set from an IBM AT. Soft configuration (ala Microport SysV/286) > is perfect, but any other configurable solution is also fine (I'm willing > to link a new kernel to change it, for example). What is *NOT* fine is > a scheme which limits you to those types which are in your BIOS. I believe > that this disqualifies Interactive (from their literature), but perhaps > that's not the whole story... > There were more than 60 beta sites for V.3/386. The beta period lasted from November 86 until the port was certified in July. While no operating system is bug-free, the V.3/386 kernel is much more reliable than many first releases. At INTERACTIVE, we do a substantial amount of development on 386 machines running 386/ix (both AT bus and Multibus). The demand from our developers is to add more 386 machines. While it is possible that some vendors may have introduced problems with their device drivers, you should find the base kernel very reliable for development purposes. Smart serial boards: this is just a device driver. 386/ix supports several multiport serial boards (of both the smart and dumb variety) and probably most other "Unix 386" vendors also support smart serial boards. Full link kit: part of the V.3/386 port. If your vendor doesn't supply it, your vendor has broken the port. In general, anyone that provides "Unix 386" will provide the standard Unix V.3/386 link kit. The one exception that I know of is AT&T. AT&T has adopted their idconfig system for their Unix release on the AT&T 6386. Big disks: This should also be no problem. INTERACTIVE developed a standard description for disk drives that is part of the certified V.3/386 port. You should be able to install any size disk. This does depend on the device driver. If your vendor did not fully implement VTOC support in his device drivers, you may have a problem. INTERACTIVE does not have a problem with any size ST-506, RLL, or ESDI disk. The INTERACTIVE literature that may have confused you mentioned the BIOS support for the disk. What is required is that the BIOS support the disk well enough to load the kernel. After the kernel is loaded, it reads the VTOC off the disk drive and reprograms the disk controller to fully support the disk. Since the AT/386 requires that the BIOS load the bootstrap, this is unavoidable. This is particularly obvious for ESDI disks. In that case, the BIOS must support ESDI controllers and disks. The bottom line is that INTERACTIVE's 386/ix does not have a problem supporting any size disk and I would expect other vendors would be able to support non-standard disks as well. > Pricing information is also appreciated, if you have it handy.... The more > complete the better! > I should note that INTERACTIVE sells 386/ix to OEMs, system integrators, VAR's, and large end users. If you don't fit into one of those categories, you would have to get 386/ix from one of our distributors. Pricing varies somewhat between distributors but 386/ix list prices are comparable with other full-service Unix/Xenix 386 vendors. Bill Lee INTERACTIVE Systems Corp. ...!ism780c!bill bill@ism780c.ism.isc.com
davidsen@steinmetz.steinmetz.UUCP (William E. Davidsen Jr) (11/24/87)
In article <361@ddsw1.UUCP> karl@ddsw1.UUCP (Karl Denninger) writes: |This has probably been hashed out a million times before, but... | |I'd like to ask the users of the net which 80386 Unix (Xenix) derivitive |they would buy. I know of the following companies producing a Unix(Xenix) |V/386 -- feel free to add to this! I am posting this to the 80386 mailing list as well |o Microport (System V/386) |o Interactive (Same) |o SCO (Xenix V/386) |o Bell Tech (?) | |We're looking for the following: | |o Speed -- Performance is very, very important. | We will be developing code on this system -- thus, the quality of | the compiler(s) also count in a big way. Comments on both the system's | speed as a whole and on the quality of the code generated by the | compiler(s) is desired. mport, bell and IS are shipping a PCC based compiler. mport *claims* that they will ship Greenhills in January, but they also told us that they shipped V/AT six weeks ago. Xenix/386 uses an optimizing compiler which will generate code for Xenix/286 and DOS as well as native code. I have played with it a bit on a Compaq and it didn't surprize me. The option to reduce common subexpressions does produce some major improvements in some code. |o Adherance to standards -- SVID is foremost, although utility-level | compatibility would also be nice. All but Xenix are based on a certified (by AT&T) port. Xenix claims to support SVID. It is much more standard that Xenix/286 as far as I can tell, although the system administration is still Xenix flavor. |o Reliability -- We cannot tolerate a product that panics or crashes. | We *need* 9600 baud communications ability, 19,200 is desired. Xenix will run two ports at 9600. More than that I don't know. Bell has drivers for smart or dumb 6 ports serial cards which go at least to 9600 (I think 19.2). They wrote the drivers for the version which was certified by AT&T. mport has had a *lot* of trouble with their serial drivers on V/AT, they tell me there's no problem on V/386. |o Must handle 'smart' serial boards (requirement above mandates it). | |o Must have a full link kit (ability to handle custom drivers *AND* tune | some kernel parameters); a section in the manuals that is worthwhile on | writing drivers for the particular varient of Unix is even more helpful. Xenix/386 has a *whole lot* of info on device drivers, including examples. I haven't seen the others enough to comment. |o MUST (repeat, MUST) handle disk drives which are not in the standard | 15-type set from an IBM AT. Soft configuration (ala Microport SysV/286) | is perfect, but any other configurable solution is also fine (I'm willing | to link a new kernel to change it, for example). What is *NOT* fine is | a scheme which limits you to those types which are in your BIOS. I believe | that this disqualifies Interactive (from their literature), but perhaps | that's not the whole story... Xenix/386 does. I'm told that V/386 does (I didn't install it). |In addition, ability to run DOS/MERGE in some form is also desired, but not |necessary (we have other systems for that if needed). mport claims that they will be shipping DOSmerge in January (we haven't gotten production versions of the product we ordered six months ago for V/AT, but they did send us buggy beta versions), and IS is shipping now. I don't know the status for Xenix. -- bill davidsen (wedu@ge-crd.arpa) {uunet | philabs | seismo}!steinmetz!crdos1!davidsen "Stupidity, like virtue, is its own reward" -me
manes@dasys1.UUCP (Steve Manes) (11/27/87)
Microport also denied for so long that there was any problem in the tty driver for 2.2 until a few weeks before shipping 2.3 out to its beta testers. Now they claim that 2.3 fixed all the tty bugs. Again, not so. The 2.3 beta I played with (supposedly the one with the fixed tty driver) is improved over 2.2 (at least you can run one 9600 baud terminal and Xmodem now) but 'cu' still loses characters if another tty line is in use. I'm still waiting for my 2.3 revision (the one that was supposedly shipped 6 weeks ago) so I don't know if the released version REALLY fixed the problem but, if not, I wonder what value their multiport serial board support is if uport can't handle even the standard two tty lines without character loss. -- +----------------------------------------------------------------------- + Steve Manes Roxy Recorders, Inc. NYC + decvax!philabs!cmcl2!hombre!magpie!manes Magpie BBS: 212-420-0527 + uunet!iuvax!bsu-cs!zoo-hq!magpie!manes 300/1200/2400
neighorn@qiclab.UUCP (Steve Neighorn) (12/01/87)
In article <2081@dasys1.UUCP> manes@dasys1.UUCP (Steve Manes) writes:
:I'm still waiting for my 2.3 revision (the one that was supposedly shipped
:6 weeks ago) so I don't know if the released version REALLY fixed the
:problem but, if not, I wonder what value their multiport serial board
:support is if uport can't handle even the standard two tty lines without
:character loss.
:--
:+ Steve Manes Roxy Recorders, Inc. NYC
Oregon Health Science University is currently running a 2.3 revision
(merge kernel) with two digiboard 8-port intelligent serial cards on
a Compaq 386 running at 16 Mhz with none of the problems previously
described in this newsgroup. This machine has been in operation for a
couple months with 16 terminals, all running at 9600 baud. Microport,
along with Digiboard, at least in this application, have gotten their
serial acts together. We will be installing V/386 and V/386 digiboard
drivers on the Compaq as soon as we can get our hands on a working
DataFlex compiler.
--
Steven C. Neighorn !tektronix!{psu-cs,reed,ogcvax}!qiclab!neighorn
Portland Public Schools "Where we train young Star Fighters to defend the
(503) 249-2000 ext 337 frontier against Xur and the Ko-dan Armada"manes@dasys1.UUCP (12/04/87)
In article <873@qiclab.UUCP> neighorn@qiclab.UUCP (Steve Neighorn) writes: >Oregon Health Science University is currently running a 2.3 revision >(merge kernel) with two digiboard 8-port intelligent serial cards on >a Compaq 386 running at 16 Mhz with none of the problems previously >described in this newsgroup. Since I'm still waiting for the 2.3 revision I paid my update contract fee for I'll have to take your word for it. But I did try the last beta revision copy of Microport V/AT 2.3 on a 10mHz AT and it continued to lock up under heavy 9600 baud load even from one terminal, let alone 16. I've only run a couple of 9600 baud terminals on uport and it worked fine for general terminal handling. The test is to run a large-block file transfer protocol, like Ymodem at 9.6k. I've never been able to successfully do this on Microport, including the last beta-rev of 2.3 I played with. My experience with Digiboard is another war story. But this all belongs in the new comp.unix.microport... -- +----------------------------------------------------------------------- + Steve Manes Roxy Recorders, Inc. NYC + decvax!philabs!cmcl2!hombre!magpie!manes Magpie BBS: 212-420-0527 + uunet!iuvax!bsu-cs!zoo-hq!magpie!manes 300/1200/2400