OK, got it working.
The crux was to throw all
Rcpp related stuff away from the parallel
context, even the Rcpp::List returned by
m_psiMinRVec[i]->init(...).
Replacing only
parInit(const Rcpp::List& seedList) with
parInit(std::vector >& made it
a "little bit more robust", but still produced
"Warn
> But your threads,
> be it pthreads or C++11 threads are not calling the RNG supplied by R,
are they?
That is correct. Just reading and writing to c++ vectors (with some non-R
deterministic computations in-between)
Best,
--
Hao Ye
[email protected]
On Mar 11, 2014, at 1:15 PM, Dirk Eddelbuettel wr
you can also use intel TBB, just depends on your preference for syntax
I suppose. I've never tried to use openMP threads, but intel tbb
works quite nicely.
example here based off of rrcov package:
https://github.com/armstrtw/rrcov/blob/master/src/covOPW.cpp#L58
Sorry, Dirk, this doesn't use Rcpp
On 11 March 2014 at 13:06, Hao Ye wrote:
| > Just accessing / setting data structures should work (but test first...),
| > RNGs is clearly asking for trouble.
|
| FWIW, I've been able to get this working using Rcpp and C++11 threads. (but
using native C++ data structures, of course)
Sure, just
> Just accessing / setting data structures should work (but test first...),
> RNGs is clearly asking for trouble.
FWIW, I've been able to get this working using Rcpp and C++11 threads. (but
using native C++ data structures, of course)
Best,
--
Hao Ye
[email protected]
On Mar 11, 2014, at 12:56 PM,
Andreas,
On 11 March 2014 at 20:30, Andreas Prescher wrote:
| Hello,
|
| I got strange behaviour when using
| OpenMp and IntegerVector (without
| OpenMp all works fine!!)
|
| From my package (Rcpp 0.10.4 used, let some unimportant
| arguments out)
|
| Rcpp::List PsiMinRContainer::parInit(cons
