HGCD_REDUCE_THRESHOLD

Torbjorn Granlund tg at gmplib.org
Sun Nov 13 00:54:55 CET 2011


nisse at lysator.liu.se (Niels Möller) writes:

  I tried some values, amd I think max_size = 10000 (or something a bit
  smaller, 5000 or 7000) and step_factor = 0.4 should give reasonable
  tuning time. When I run ./tuneup -t here, I get
  
  #define HGCD_REDUCE_THRESHOLD            ...
  sizes 30 to 9748 total 162 measurements
  #define HGCD_REDUCE_THRESHOLD             2205
  
  I'm considering checking in that change.

Perhaps 0.4 is a bit too coarse?
  
  BTW, it would be nice if tuneup -t also showed the total running time
  for each (or most) tuned threshold.
  
Funny, I was thinking of that too.  And that you should fix it.  :-)

-- 
Torbjörn


More information about the gmp-devel mailing list