[Rcpp-devel] Stack imbalance warning when using Rcpp and OpenMP

[Michael Braun]

Hi.  Just one last clarifying question on this issue before I dive back in.

Suppose I declared a new Rcpp::List object in my C++ code, and copied the list elements from either the SEXP or the original Rcpp::List.  Since the new memory is allocated in C++, would I still have the same problem because of the way Rcpp allocated the memory?  Or would the copy be thread-safe?

Similarly, what if I were to create an STL container of Rcpp::Lists, and operate on each element of the container in parallel?  Same problem?

From your helpful responses, it seems like the best alternative is to explicitly copy the contents of each SEXP in the list to a totally non-Rcpp object. I'm just wondering if keeping some of the data in the original classes might still work.

And finally, since I am still relatively new to C++, are there any standard classes that might make more sense than others?  I'm considering an STL vector of either Eigen or Armadillo matrices, for example.  Good idea, or bad?

Thanks again,

Michael


On Jul 27, 2011, at 3:55 PM, Dirk Eddelbuettel wrote:

> 
> On 27 July 2011 at 15:04, Michael Braun wrote:
> | It is taking the SEXP as the argument.  If R is the Rcpp::List argument, and i is the index of the parallelized loop, then:
> | 
> | double y = as<double>(R(i));
> | double res = func(y);
> | 
> | does work.  So far, as long as the Rcpp stuff happens in the loop directly (same function), it works.  But if I call another function from the parallelized loop, and do some Rcpp work in that function, that's when I get problems.
> 
> Yes, you cannot 'multithread' R just like that. R is single-threaded and
> (AFAIK) not reentrant.  Just because you'd like it to does not make it
> capable of multi-threaded operations.
> 
> | So while it is great that this works on this simple example, my real-world application is much more complicated (e.g., passing lists of lists), and this kind of workaround doesn't help at all.  But let's get the simple case fixed first.
> 
> You may have to copy your data out of R containers.  Or seperate the R
> processes: if you shield different R sessions, you can have several. See
> Rserve, multicore, Rmpi, ...
> 
> | Are you aware of anyone at all who has had success combining Rcpp and OpenMP?
> 
> If you are careful about your entry and exit from blocks that contain
> multithreaded code, you can use OpenMP -- see the aforementioned example in
> the Rcpp sources.
> 
> Hope this helps,  Dirk
> 
> | On Jul 27, 2011, at 2:55 PM, Davor Cubranic wrote:
> | 
> | > On 2011-07-27, at 10:33 AM, Michael Braun wrote:
> | > 
> | >> I want to avoid making a copy of the entire list (my real-world application might have lots of different elements for each element of the "master" list).  But what works fine is if, within the loop, I do the Rcpp-to-double conversion, and then call a version of the function that takes a double as an argument, instead of the SEXP.  So I think there is something going on with passing the SEXP to func, and then using Rcpp::as to convert the SEXP to double.
> | >> 
> | >> Is this close to what you are asking?
> | > 
> | > Yep. I'm still surprised that Rcpp:as works (I thought it would also take a SEXP as an argument), but if it works, then great.
> | > 
> | > Davor
> | 
> | -------------------------------------------
> | Michael Braun
> | Associate Professor of Management Science (Marketing Group)
> | MIT Sloan School of Management
> | 100 Main St.., E62-535
> | Cambridge, MA 02139
> | braunm at mit.edu
> | 617-253-3436
> | 
> | 
> 
> -- 
> Gauss once played himself in a zero-sum game and won $50.
>                      -- #11 at http://www.gaussfacts.com

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 1844 bytes
Desc: not available
URL: <http://lists.r-forge.r-project.org/pipermail/rcpp-devel/attachments/20110819/af168f7e/attachment.bin>