[SciPy-user] Solver with n-dimentional steps

Yosef Meller yosefmel@post.tau.ac...
Mon Jun 16 05:53:15 CDT 2008


On Sunday 15 June 2008 16:55:47 Hoyt Koepke wrote:
> I'm suggesting this partly to satisfy my own curiosity, so ignore this
> if it is difficult.   Since you are running with so many dimensions,
> you might want to try minimizing the square of your function using
> fmin_l_bfgs_b.  lbfgs is written for higher dimensions, and I've heard
> it can give incredible speed improvements over many other methods.
> I'm curious if it would help in your case.

Thanks for the tip. After trying half a day to figure out why a jacobian that 
should take 1GiB (8 byte floats) clogs the memory in a 2GiB machine, I 
decided to go for  fmin_l_bfgs_b(). I can't tell you if it improves the 
speed, but the low memory requirements make it work for me, as opposed to 
fsolve().


More information about the SciPy-user mailing list