Project

General

Profile

Slug #1369

RandomUnimodularMat is slow with many iters

Added by John Abbott over 4 years ago. Updated over 4 years ago.

Status:
In Progress
Priority:
Low
Assignee:
-
Category:
Improving
Target version:
CoCoALib-1.0
Start date:
25 Nov 2019
Due date:
% Done:

10%

Estimated time:
Spent time:
0.75 h

Description

RandomUnimodularMat becomes too slow when number of iters is high.

Try a divide-and-conquer approach.

Related issues

Related to CoCoALib - Feature #955: RandomUnimodularMat - random unimodular matrix (of integers)Closed2016-10-27

History

#1 Updated by John Abbott over 4 years ago

  • Related to Feature #955: RandomUnimodularMat - random unimodular matrix (of integers) added

#2 Updated by John Abbott over 4 years ago

RandomUnimodularMat(QQ,4,1000000) takes too long. It is faster to do product([RandomUnimodularMat(QQ,4,10000) | j in 1..100])

Divide-and-conquer makes sense; not sure when it should switch over to simple linear approach.

#3 Updated by John Abbott over 4 years ago

  • Status changed from New to In Progress
  • % Done changed from 0 to 10

First impl is quite disappointing; the overhead/cost of matrix mult is embarrassingly high.
Should do recursion based on the internal vector< vector<BigInt> > structure then convert to matrix only at the end.
Not tonight, Josephine!

#4 Updated by John Abbott over 4 years ago

  • Target version changed from CoCoALib-0.99700 to CoCoALib-1.0

Also available in: Atom PDF