The statement in Chris' article, Memory-mapped files can be much larger than system RAM, yet they can be accessed without page thrashing.
needs further investigation in 64-bit J. You can write a dummy 10G file using linux command, dd if=/dev/zero of=foo bs=1M count=10000 and play with it (increase file size if you get more than 10G RAM). Пнд, 01 Авг 2011, Steven Taylor писал(а): > Hi Bill, > > so in J does the OS still manage pulling parts of the file in / out of > memory? > > optimistic case: > Does 'size of viewport' mean that J takes x amount of the addressable > memory, yet the OS only loads in parts relevant to the current operation? > i.e. so long as you stay away from some safe threshold (say 1e9 * 0.66), you > *should* be okay? In a 32 bit OS it's about 4 GB (1e_9 * 2^32 GB)? > > Thanks for the info Bill, Raul, and Lettow. > > re: JDB. I'll come back to this when I get a spare moment. I'm not sure I > want to tackle a new database just yet :) > > re: vector article, is that available on the vector.org.uk website > somewhere, or in the printed periodical? > > thanks, > -Steven > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm -- regards, ==================================================== GPG key 1024D/4434BAB3 2008-08-24 gpg --keyserver subkeys.pgp.net --recv-keys 4434BAB3 ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
