Greetings, and thanks for pointing this out! As you might suspect, this is a weakness in the hash function algorithm for this set of keys, which is currently generating too many collisions.
I've committed a fix into the stable and experimental branches which you might like to look at. Am also updating the appropriate Debian packages. Take care, Barton Willis <[email protected]> writes: > Assign and access a 40,000 member hashtable > > GCL: 100.6400 seconds > CCL: 0.7800 seconds > > This has been pointed out before, I think. The code: > > (%i1) showtime : true$ > Evaluation took 0.0000 seconds (0.0000 elapsed) > (%i2) block([use_fast_arrays : true, n : 200], > (for i : 1 thru n do for j : 1 thru n do a[i,j] : i+j, > s : 0, > for i : 1 thru n do for j : 1 thru n do s : s + a[i,j])); > Evaluation took 100.6400 seconds (100.6400 elapsed) > (%o2) done > > (%i3) s; > Evaluation took 0.0000 seconds (0.0000 elapsed) > (%o3) 8040000 > (%i4) block([use_fast_arrays : false, n : 200], > (for i : 1 thru n do for j : 1 thru n do b[i,j] : i+j, > s : 0, > for i : 1 thru n do for j : 1 thru n do s : s + b[i,j])); > Evaluation took 0.7700 seconds (0.7700 elapsed) > (%o4) done > (%i5) s; > Evaluation took 0.0000 seconds (0.0000 elapsed) > (%o5) 8040000 > > > With CCL: > > (%i1) showtime : true$ > Evaluation took 0.0000 seconds (0.0000 elapsed) > (%i2) block([use_fast_arrays : true, n : 200], > (for i : 1 thru n do for j : 1 thru n do a[i,j] : i+j, > s : 0, > for i : 1 thru n do for j : 1 thru n do s : s + a[i,j])); > Evaluation took 0.7800 seconds (0.7800 elapsed) > (%o2) done > (%i3) s; > Evaluation took 0.0000 seconds (0.0000 elapsed) > (%o3) 8040000 > (%i4) block([use_fast_arrays : false, n : 200], > (for i : 1 thru n do for j : 1 thru n do b[i,j] : i+j, > s : 0, > for i : 1 thru n do for j : 1 thru n do s : s + b[i,j])); > Evaluation took 0.8892 seconds (0.8900 elapsed) > (%o4) done > (%i5) s; > Evaluation took 0.0000 seconds (0.0000 elapsed) > (%o5) 8040000 > (%i6) > > > > > --Barton > > _______________________________________________ > Maxima mailing list > [email protected] > http://www.math.utexas.edu/mailman/listinfo/maxima > > > > -- Camm Maguire [email protected] ========================================================================== "The earth is but one country, and mankind its citizens." -- Baha'u'llah _______________________________________________ Gcl-devel mailing list [email protected] https://lists.gnu.org/mailman/listinfo/gcl-devel
