[Default] On 4 May 2017 15:53:17 -0700, in bit.listserv.ibm-main [email protected] (Steve Thompson) wrote:
>On 05/04/2017 06:08 PM, Clark Morris wrote: ><SNIPPAGE> >> Since this is old Fortran, I suspect that the given simulation can >> only use one core. It also may be limited in the memory it can use or >> does use. Having a good optimizing compiler would help and one may or >> may not be available for the box used. There may be code which was >> optimal for the environment in existence when written but >> counter-productive in the current environment. I know that >> optimization that I did for COBOL VS was in some ways >> counter-productive in VS COBOL II 1.4 and later. >> >> Clark Morris > >Would this be "OLD" FORTRAN because someone knows that FORTRAN is >an "old" language, or would this be "OLD" FORTRAN II? Knowing the resistance to change by some organizations I could believe Fortran IV and maybe II upgraded to the extent needed to run on the most current compiler. > >I'm thinking, to even get this to work on the "massively >parallel" system it is running on, that one is having to use the >compiler for that O/S. > >If that is the case, and they can get their version of FORTRAN 77 >(?) to process the source, then this code is probably already >doing multi-tasking for as much as the compiler can determine can >be done. Will Fortran 77 automatically determine code that can be multi-tasked? From what little I have read, I thought programs had to be architected to take advantage of multi-tasking and that this was a major effort. > >Now, if one knows the system, and one knows the various >implementer extensions, one probably could make this run faster. > >This has one wondering if this code was originally running on an >OLD Cray or something similar. > >But then again we all know that "OLD" COBOL 6 is running a lot of >"OLD" COBOL source to make lots of money out of obsolete programs >that process obsolete transactions on "OLD" obsolete Mainframes >to make a lot of obsolete Bit Coin... Wait! The payroll programs I wrote were replaced by a package. The marketing programs should have been replaced long before I left the company in 1991 because they really were not adequate for the business. In both cases some of the programs were a nightmare including the use of the ALTER verb to fit in small (16K to start with growing to 24K) DOS partitions. The main thing that holds many companies back from replacing systems is that no one knows exactly what they do and companies aren't willing to spend the money to either find out or even provide adequate testing capability. I know that as a systems programmer I had no way of running any application suite to make sure a new release didn't blow it up. As a contract programmer figuring out how to test programs took more time than modifying them and writing the changes. Hopefully things have gotten better after I retired and maybe I just picked companies that were not as well run as they should be. On the other hand, when I was working on the year 2000 project at one company they put in a far more formal process of testing. Many systems that are run in environment X might well be more effectively run in environment Y. The mainframe is better for some that are currently on distributed and the reverse is true. > >What? > >Yeah, > >Regards, >Steve Thompson > >ps. Sorry Clark, your posting was just begging to be used and abused. Glad to provide a good diving board. Clark Morris > >---------------------------------------------------------------------- >For IBM-MAIN subscribe / signoff / archive access instructions, >send email to [email protected] with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
