Ahh, Jay's comments about the computer at Ga. Tech brings back many memories
of when I was at Ga. Tech (undergraduate '66 to '70 & graduate '75 to '76).
I, too, struggled thru my Fortran class until a friend of mine told me about
the "parallel processing" trick to vastly speed up the time to achieve the
optimum solution. Turn-around on those punch cards was still terrible. We
would
make copies of the card stacks and change one or more cards with different
commands and we would end up with 10 or more stacks (have to keep track of what
was different in each one) and run ALL of the them at one time, throw away the
stacks with bad results and move forward with the good (fix the fewest
mistakes) stacks and thus eventually (several runs later) arrive at just one
stack and
printout. This process saved about 80% of the time to solution over the
sequential process that Jay suffered through including all the
copying/changing/managing required of the parallel approach. Saved my butt in
that class. I
still sometimes use this approach in solving problems quicker in my business
environment.
Grizzled old veteran who has been autrocrossing since 1970.
Jim Murphy
1984 DareDevil F500 F Mod #40
|