Frederick Ford asks about GAP on super/parallel computers. I am sorry
that from Aachen at present we cannot give much advice on his
questions, we have not tried any of the systems that are mentioned
here. I think GAP is used on an RS-6000 e.g. at Essen , perhaps
Gerhard Schneider (formerly Essen, now Karlsruhe) can give advice on
IBMs generally. I seem to remember that the question of GAP on a
supercomputer has been mentioned in the connection of measuring
GAPstones, but I have not kept that correspondence. Perhaps somebody
from the GAP-forum has some experience. Also Martin Schoenert may have
some information from private correspondence, but he will return from
holidays at the beginning of March only.
Generally I would not expect too much gain from a parallel machine
since there is nothing in GAP specially supporting parallel computing,
but this is pure (maybe poor) guessing. I you dare to try, we would be
interested to hear of the outcome.