Databases
StackOverflow2013
Postsdbo
POWhy is foreach() %do% sometimes slower than for?
 

Note that there are some explanatory texts on larger screens.

plurals
  1. POWhy is foreach() %do% sometimes slower than for?
    primarykey
    data
    text
    <p>I'm playing around with parallellization in R for the first time. As a first toy example, I tried</p> <pre><code>library(doMC) registerDoMC() B&lt;-10000 myFunc&lt;-function() { for(i in 1:B) sqrt(i) } myFunc2&lt;-function() { foreach(i = 1:B) %do% sqrt(i) } myParFunc&lt;-function() { foreach(i = 1:B) %dopar% sqrt(i) } </code></pre> <p>I know that <code>sqrt()</code> executes too fast for parallellization to matter, but what I didn't expect was that <code>foreach() %do%</code> would be slower than <code>for()</code>:</p> <pre><code>&gt; system.time(myFunc()) user system elapsed 0.004 0.000 0.005 &gt; system.time(myFunc2()) user system elapsed 6.756 0.000 6.759 &gt; system.time(myParFunc()) user system elapsed 6.140 0.524 6.096 </code></pre> <p>In most examples that I've seen, <code>foreach() %dopar%</code> is compared to <code>foreach() %do%</code> rather than <code>for()</code>. Since <code>foreach() %do%</code> was much slower than <code>for()</code> in my toy example, I'm now a bit confused. Somehow, I thought that these were equivalent ways of constructing for-loops. What is the difference? Are they ever equivalent? Is <code>foreach() %do%</code> always slower?</p> <p>UPDATE: Following @Peter Fines answer, I update <code>myFunc</code> as follows:</p> <pre><code> a&lt;-rep(NA,B) myFunc&lt;-function() { for(i in 1:B) a[i]&lt;-sqrt(i) } </code></pre> <p>This makes <code>for()</code> a bit slower, but not much:</p> <pre><code>&gt; system.time(myFunc()) user system elapsed 0.036 0.000 0.035 &gt; system.time(myFunc2()) user system elapsed 6.380 0.000 6.385 </code></pre>
    singulars
    1. PO
      singulars
      1. PTAnswer
    1. This table or related slice is empty.
    1. PTQuestion
    1. USMånsT
    1. USMånsT
    plurals
    1. PL
      singulars
      1. LTLinked
    2. PL
      singulars
      1. LTLinked
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. PO
      singulars
      1. PTAnswer
    1. VO
      singulars
      1. POWhy is foreach() %do% sometimes slower than for?
      1. This table or related slice is empty.
      1. VTUpMod
    2. VO
      singulars
      1. POWhy is foreach() %do% sometimes slower than for?
      1. USBenBarnes
      1. VTFavorite
    3. VO
      singulars
      1. POWhy is foreach() %do% sometimes slower than for?
      1. This table or related slice is empty.
      1. VTUpMod
    1. COSee also this question: http://stackoverflow.com/questions/5007458/problems-using-foreach-parallelization and this one: http://stackoverflow.com/questions/5012804/mpi-parallelization-using-snow-is-slow
      singulars
      1. POWhy is foreach() %do% sometimes slower than for?
      1. USCharlie
    2. COThanks @Charlie, the answers to those questions were very helpful to what I'm trying to do after I'm done with my toy example! :) I'm still not sure that I understand why `foreach` needs so much more time when using the `%do%` option though.
      singulars
      1. POWhy is foreach() %do% sometimes slower than for?
      1. USMånsT
    3. COA big part of it is that %do% has to parcel out the pieces/assignments, send them to the processors, then rejoin everything at the end as appropriate. These steps require organizational time that the unparallelized version doesn't.
      singulars
      1. POWhy is foreach() %do% sometimes slower than for?
      1. USCharlie
 

Querying!

 
Guidance
A row detail

Detail views are divided into sections. All the information in the data section comes from columns in the selected row. The other sections display data from other, related rows.

Related data can be related in a to-one or a to-many fashion. Captions of data related in a to-many fashion link to a list view showing a filtered view of the table.

Try moving around until you find a non-empty to-many entry and click on the label to get to one. You can move back to the root by clicking on the database name in the header.

SQuiL has stopped working due to an internal error.

If you are curious you may find further information in the browser console, which is accessible through the devtools (F12).

Reload