Frequently Asked Questions
Work under progress
- Does it work?
- I realize I made a mistake, how can I stop it?
- Can I print any output from the parallelized loop to the screen?
- Is there a way to know how far my calculation has progressed?
- Is it free?
- Can I contribute?
- Does it only work for desktop computers?
- How can I build the package from the source files?
1.- Does it work?
Oh yeah. Nicely.
Anyway is still in beta version. Some cases has been reported where the calculation never ends because of a faulty worker.
We are working hard to isolate these bugs and fix them.
2.- I realize I made a mistake, how can I stop it?
Yes, and no. The right way should be to type CTRL-C as usual in R to cancel the ongoing calculation.
This feature is not yet implemented. Right now, if you type CTRL-C you will recover the R prompt, but in the background
the workers will keep 'working' until the end. You have to stop them killing/stoping the processes.
3.- Can I print any output from the parallelized loop to the screen?
Yes, you can. At your own risk. The loop is splitted and processed by the workers, but during processing, the order in which each iteration
is run by each worker is unknown and printed information can appear at any unexpected order.
4.- Is there a way to know how far my calculation has progressed?
Not yet. For sure, less than running it sequentially. We plan to provided a text-based progress bar or similar.
Also an 'Estimation Time Arrival' is planned for next distributed version. Please, stay tuned.
5.- Is it free?
Yap. License changed to GPL.
You can contact us if the license does not fit to your needs.
6.- Can I contribute?
Of course you can! We will very happy to receive any suggestions, comments, usage experience or even development collaborations.
7.- Does it only work for desktop computers?
And servers! It works in any computer with the right operating system (Windows XP and Linux), architecture (x86) and R version (2.6.x). For example we are using it with a linux server with 16 processors and 128 GB with impressive results.
8.- How can I build the package from the source files?
(not finished)
In linux: R CMD build --binary rparallel
In Windows: R CMD BUILD --binary rparallel
Additional packages to compile with R in Linux: r-base-dev in Ubuntu, R-devel in Fedora.