<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix">On 11/6/13 6:57 PM, Douglas Bates
wrote:<br>
</div>
<blockquote
cite="mid:CAO7JsnQP6zaOP8g5L681Q-AHCCROBKb1OWM3PnGVGAbWZn2Wpg@mail.gmail.com"
type="cite">
<div dir="ltr">By default Eigen does not use BLAS, which can be
good or bad, depending on the situation. I notice that the
second largest total time is spent in t.default which may mean
that you are using an operation like
<div>
<br>
</div>
<div>t(X) %*% X</div>
<div><br>
</div>
<div>If so, you can save yourself time by using the crossprod or
tcrossprod functions. For example, the expression above is
more cleanly written as</div>
<div><br>
</div>
<div>crossprod(X)</div>
<div><br>
</div>
<div>or, for t(X) %*% Y,</div>
<div><br>
</div>
<div>crossprod(X, Y)<br>
<br>
Eigen or Armadillo could help you avoid making unnecessary
copied but if your calculation does end up being dominated by
matrix multiplications you can't expect to gain much speed
relative to R. You may want to check the type of BLAS you are
using. For Intel processors MKL is generally the fastest (but
proprietary) with OpenBLAS in second place.<br>
<br>
I see that I am, as often happens, giving you similar advice
to Dirk's response.</div>
</div>
</blockquote>
t(X) %*% Y is indeed a very frequent operation. Thanks for pointing
out the crossprod function, I was unaware of its existence.<br>
<br>
<br>
Xavier<br>
<br>
<blockquote
cite="mid:CAO7JsnQP6zaOP8g5L681Q-AHCCROBKb1OWM3PnGVGAbWZn2Wpg@mail.gmail.com"
type="cite">
<div class="gmail_extra"><br>
<br>
<div class="gmail_quote">On Wed, Nov 6, 2013 at 11:35 AM, Xavier
Robin <span dir="ltr"><<a moz-do-not-send="true"
href="mailto:xavier@cbs.dtu.dk" target="_blank">xavier@cbs.dtu.dk</a>></span>
wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">Hi,<br>
<br>
I have a pure-R code that spends most of the time performing
vector and matrix operations, as shown by the summary of
Rprof:<br>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">
self.time self.pct total.time total.pct<br>
"%*%" 903.24 77.67 903.24 77.67<br>
"t.default" 76.26 6.56 76.26 6.56<br>
"-" 36.60 3.15 36.60 3.15<br>
"+" 24.44 2.10 24.44 2.10<br>
"/" 24.22 2.08 24.22 2.08<br>
"exp" 20.26 1.74 20.26 1.74<br>
"predict.myClass" 17.68 1.52 503.82 43.32<br>
"*" 11.90 1.02 11.90 1.02<br>
"t" 9.38 0.81 811.94 69.82<br>
"update.batch" 8.04 0.69 654.68 56.30<br>
...<br>
</blockquote>
So mostly matrix %*% matrix multiplications, transpositions,
vector +-/* matrix operations and exponentiations,
representing >95% of the computation time.<br>
I have very few loops and if/else blocks.<br>
<br>
I want to speed up this code, and I am considering
reimplementing it (or part of it) with RcppEigen or
RcppArmadillo.<br>
<br>
However, I read that both Eigen and Amarillo use the
underlying BLAS, like R.<br>
My question is, can I expect any significant speed-up from
an Rcpp re-implementation in this case, given it is already
mostly matrix algebra (which are supposed to be pretty
efficient in R)?<br>
<br>
Thanks,<br>
Xavier<span class="HOEnZb"><font color="#888888"><br>
<br>
-- <br>
Xavier Robin, PhD<br>
Cellular Signal Integration Group (C-SIG) - <a
moz-do-not-send="true"
href="http://www.lindinglab.org" target="_blank">http://www.lindinglab.org</a><br>
Center for Biological Sequence Analysis (CBS) - <a
moz-do-not-send="true" href="http://www.cbs.dtu.dk"
target="_blank">http://www.cbs.dtu.dk</a><br>
Department of Systems Biology - Technical University of
Denmark (DTU)<br>
Anker Engelundsvej, Building 301, DK-2800 Lyngby,
DENMARK.<br>
<br>
_______________________________________________<br>
Rcpp-devel mailing list<br>
<a moz-do-not-send="true"
href="mailto:Rcpp-devel@lists.r-forge.r-project.org"
target="_blank">Rcpp-devel@lists.r-forge.r-project.org</a><br>
<a moz-do-not-send="true"
href="https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/rcpp-devel"
target="_blank">https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/rcpp-devel</a><br>
</font></span></blockquote>
</div>
<br>
</div>
</blockquote>
<br>
<br>
<pre class="moz-signature" cols="72">--
Xavier Robin, PhD
Cellular Signal Integration Group (C-SIG) - <a class="moz-txt-link-freetext" href="http://www.lindinglab.org">http://www.lindinglab.org</a>
Center for Biological Sequence Analysis (CBS) - <a class="moz-txt-link-freetext" href="http://www.cbs.dtu.dk">http://www.cbs.dtu.dk</a>
Department of Systems Biology - Technical University of Denmark (DTU)
Anker Engelundsvej, Building 301, DK-2800 Lyngby, DENMARK.</pre>
</body>
</html>