Jared Hoberock
Jared Hoberock
Since InputIterators can't be dereferenced multiple times we shouldn't implement algorithms like copy_if(first, last, output, pred) with copy_if(first, last, first, output, pred). In general the unit tests should check InputIterator...
We should not require types to have default constructors unless the algorithm truly requires them. Eventually we should test our standard algorithms with user-defined types that do not provide a...
Reported by andrew.corrigan, Nov 2, 2010 Is it possible to add "#pragma ivdep" [1] on the line following any "#pragma omp parallel for" directives? This can facilitate vectorization with the...
We're missing these algorithms (which look hard); partial_sort partial_sort_copy includes nth_element Forwarded from http://code.google.com/p/thrust/issues/detail?id=423
This is a set op that we forgot I guess Forwarded from http://code.google.com/p/thrust/issues/detail?id=411
maximum & minimum & absolute_value et al don't belong in thrust/functional.h but they are useful because algorithms can't take fcn ptrs yet Thrust vectors should always fire constructors but a...
z_iter[(i,j,k)] could potentially make sense if the tuple is the right size and the element types are each convertible to zip_iterator::difference_type Comment 1 by andrew.corrigan, Sep 16, 2011 I implemented...
standardize on style Forwarded from http://code.google.com/p/thrust/issues/detail?id=151
Ideally, we want a human-readable compiler error. When we have static_assert, we can guard just the function bodies and do something like static_assert(_OPENMP, "Did you forget -fopenmp?") Forwarded from http://code.google.com/p/thrust/issues/detail?id=94
Reported by repalov, Aug 3, 2010 Some execution paths in Thrust can throw exception in destructor. It is lead to call terminate() if we already processing some exception (for example...