---Randy MacDonald wrote: > Interesting numbers. My next questions would be: how modern are the > drives, and how much RAM is on the machine? Also, how long > did the OP's > 'slow' process take? Did an actual copy occur? I could see *nix > systems doing some inode magic.
Numbers were obtained on a Toshiba Tecra M9 laptop, bit less than a year old, running Vista SP1 with 2GB of RAM. The fcopynew verb from the files.ijs script does a fread & then an fwrite which suggests to me that an actual copy was done. What numbers did you get on your machine? > Sherlock, Ric wrote: > > ---Randy MacDonald wrote: > > > >> By copy, I don not mean duplicating a reference. I mean having two > >> identical sets of bytes where there was one before. What the > >> OP seemed > >> to be talking about was a filter of some sort. > >> > >> I based my estimate on 134MB being a 2 hour mp3, or a > >> 6-minute video. I > >> just thought the data transfer speed of a hard drive would > work out to > >> about a minute, or 4MB/sec, since both to and from storage > are what we > >> are talking about here. (note to self: Google 'hard drive > >> throughput' > >> for more info) > >> > >> > > > > The following seems to suggest that 1 minute would be excessive? > > > > 6!:2 '(134125010$''abcdefg'') fwrite jpath ''~temp/testsz.txt''' > > 6.27771190193 > > 6!:2 '(jpath ''~temp/testsz2.txt'') fcopynew jpath > ''~temp/testsz.txt''' > > 4.60973992505 ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
