Do you avoid monadic #: for some reason? a=: 1e6 ?@$ 256
ts '2 2 2 2 2 2 2 2 #: a' 0.241633 6.71107e7 ts '#: a' 0.0183598 8.38989e6 On Wed, Sep 11, 2013 at 6:56 PM, Pascal Jasmin <[email protected]>wrote: > I noticed that there is no special code for any of the following: (J7-64) > > a =. ? 1e6 # 256 > ts '2 2 2 2 2 2 2 2 #: a' > > 0.124369 6.71107e7 > ts '4 2 4 2 4 2 #: a' > 0.0943141 6.71105e7 > a2 =.? 1e6 # 256^2 > ts '8 8 4 2 4 2 4 2 #: a2' > 0.144685 6.71107e7 > a3 =.? 1e6 # 1-~256^4 > a2 =.? 1e6 # 1-~256^2 > ts '8 8 4 2 4 2 4 2 #: a2' > 0.141351 6.71107e7 > ts '256 256 8 8 4 2 4 2 4 2 #: a3' > 0.247149 1.3422e8 > > ts '512 512 128 128 #: a3' > 0.0837122 3.35561e7 > ts '256 256 256 256 #: a3' > 0.0805493 3.35561e7 > > Is there a way to do bit unpacking more efficiently (space wise)? I > notice the huffman encoding lab is faster and smaller even though it seems > to be doing much more work. > > t=: 'the quick brown fox jumps over the lazy dog ' > > x=: 1e6 $ t > > > ts 'HA hencode x' > 0.0267345 1.67822e7 > ts 'SB hdecode y' > 0.0491284 1.78308e7 > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm > ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
