mahout-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ted Dunning <ted.dunn...@gmail.com>
Subject Re: Making Very Large-Scale Linear Algebraic Computations Possible Via Randomization
Date Tue, 29 Sep 2009 21:32:11 GMT
Probably better to save it.  Probably also depends on the problem set.  If A
has fewer non-zero elements, then recomputing y may be faster than
retrieving a saved copy.  Even if y and A have the same number of of
non-zero elements, then recomputing y might be faster because you save the
writing cost.

But I really don't know which way this should go in the end.

On Mon, Sep 28, 2009 at 11:49 PM, Jake Mannix <jake.mannix@gmail.com> wrote:

> This is true, but is recomputing Y on the second pass worth not saving it?




-- 
Ted Dunning, CTO
DeepDyve

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message