cchen@sbcs.sunysb.edu (Chyouhwa Chen) (09/23/89)
Hi, all, I just ordered a 100K lines program written in VMS fortran for my research. I would like to port it to the Sun3 environment. Besides being a more familiar environment to me, the Suns have Tektronics 4014 emulator that the program needs. ( We don't have Tek terminals.) Has anybody attempted something like this? Any advice for me before I make a fool of myself? I'm sorry I have not been following the discussions in this group closely. It never occurs to me that I will be dealing with fortran again. Thanks for any help you might offer. - Chyouhwa cchen@sbcs.sunysb.edu -------------------------
thorson@typhoon.atmos.colostate.edu (Bill Thorson) (09/25/89)
In article <3542@sbcs.sunysb.edu> cchen@sbcs.sunysb.edu (Chyouhwa Chen) writes: >I just ordered a 100K lines program written in VMS fortran for my >research. I would like to port it to the Sun3 environment. I thought all Sun's had basically the same FORTRAN compilers. Our Sun's FORTRAN came with a translator for VMS FORTRAN. The translator is called f77cvt. The way our compiler works is that if the file name ends with .vf or .for it is assumed to be VMS FORTRAN and if run throught the translator. >- Chyouhwa > cchen@sbcs.sunysb.edu Bill Thorson thorson.atmos.colostate.edu
sysruth@helios.physics.utoronto.ca (Ruth Milner) (09/27/89)
In article <2711@ccncsu.ColoState.EDU> thorson@typhoon.typhoon.atmos.colostate.edu (Bill Thorson) writes: >In article <3542@sbcs.sunysb.edu> cchen@sbcs.sunysb.edu (Chyouhwa Chen) writes: > >>I just ordered a 100K lines program written in VMS fortran for my >>research. I would like to port it to the Sun3 environment. > >I thought all Sun's had basically the same FORTRAN compilers. Our Sun's >FORTRAN came with a translator for VMS FORTRAN. The translator is called >f77cvt. The way our compiler works is that if the file name ends with >.vf or .for it is assumed to be VMS FORTRAN and if run throught the >translator. You should be aware that not all VMS extensions are supported. For example, a lot of the OPEN options are not supported under UNIX because they relate to record structure, which UNIX basically has no concept of. And, of course, if your 100K-line package has any system calls in it, you will have to either replace those calls with corresponding UNIX versions (if there are any) or write your own routines. Also, a number of simple VMS extensions were built right into f77 itself; things like END DO and DO WHILE etc. If the Fortran is not heavily VMS-dependent, you may be able to compile with just this. If you don't need it, don't run f77cvt. Although it handles many of the other things which were too big to write into f77, it has also been known to make some unnecessary changes. I don't know about Fortran 1.2, but under 1.0 it would, for example, do things like change quoted character strings in WRITE statements into Hollerith (ack). Neither necessary nor desirable. If your package was written for VMS, it is likely to be very difficult to port. In my experience, vendors writing specifically for VMS Fortran tend to make heavy use of its extensions, particularly in file formats. If it is essentially mathematical in nature, though, you may be able to get away with it since standard Fortran 77 covers what is generally needed. -- Ruth Milner UUCP - {uunet,pyramid}!utai!helios.physics!sysruth Systems Manager BITNET - sysruth@utorphys U. of Toronto INTERNET - sysruth@helios.physics.toronto.edu Physics/Astronomy/CITA Computing Consortium
sysruth@helios.physics.utoronto.ca (Ruth Milner) (09/28/89)
It has been pointed out to me that the wording in my previous followup was misleading: > In article <1989Sep26.210148.7903@helios.physics.utoronto.ca> you write: >> >>Also, a number of simple VMS extensions were built right into f77 itself; >>things like END DO and DO WHILE etc. If the Fortran is not heavily >... >>If it is essentially mathematical in nature, though, you may be able to >>get away with it since standard Fortran 77 covers what is generally >>needed. > >These seem to imply that END DO and DO WHILE are part of the Fortran 77 >standard. That is false; they are VMS extensions, and vanilla f77 on >UNIX systems has not historically supported them. > ... > -P. >-- >************************f*u*cn*rd*ths*u*cn*gt*a*gd*jb************************** >Peter S. Shenkin, Department of Chemistry, Barnard College, New York, NY 10027 >(212)854-1418 shenkin@cunixc.cc.columbia.edu(Internet) shenkin@cunixc(Bitnet) This is of course right. I apologize for any ambiguity. What I was actually trying to point out in the first part was that *Sun's* Fortran with VMS extensions has these particular items built right into the f77 program, making the use of f77cvt often unnecessary. I did not mean to imply that it was part of everyone else's UNIX f77. (We are after all talking about Sun 3's in this particular discussion, not UNIX systems in general - or at least that's what I thought). In the second part, I simply meant that mathematical programs often do not need to make use of VMS extensions in order to get the job done efficiently, so that if that is the kind of thing this particular package does, it may not be an issue. If it was written for VMS, though, there is almost certainly something in 100,000 lines that even Sun's f77cvt can't swallow. Again, apologies if there were other people who read it that way. -- Ruth Milner UUCP - {uunet,pyramid}!utai!helios.physics!sysruth Systems Manager BITNET - sysruth@utorphys U. of Toronto INTERNET - sysruth@helios.physics.toronto.edu Physics/Astronomy/CITA Computing Consortium