Joseph Schuchart

Thanks for the enlightenment! Then maybe `accumulate_impl` should not reside in `Accumulate.h` since the semantics are different? Who owns the code?

I know. Let me rephrase: Who will fix it? ;)

Reopening this issue after accidentally looking at the code for `dash::transform`: The way I understand the semantics of `dash::transform` is that it performs pair-wise operations on two input ranges into...

Right, got the docs wrong. Sorry for that.

I have to reopen this once again because the implementation of dash::accumulate is still broken, which was the actual reason for this issue. Please see my first paragraph in the...

@fmoessbauer Can you reproduce the MPI-IO problem with a small example or describe what the code is doing so we can file a bug report?

@fuchsto @rkowalewski @fuerlinger This is a question to you: what is the iteration order of a global iterator through a tiled matrix? My understanding is that it should still go...

@rkowalewski Thanks for the clarification! I was wrongly assuming that global iterators follow the row-iteration order. We should definitely add this capability, maybe as a specialization of `GlobIter`? The use...

@Goon I cannot seem to reproduce this issue: ```C #include int main() { dash::SizeSpec dash_size_spec = dash::SizeSpec(); return 0; } ``` ``` $ ~/opt/dash-0.4.0/bin/dash-mpicxx test_sizespec.cc $ ``` What compiler are...

@fuchsto When can we expect #410 to land? Local-to-global clearly is broken (or incomplete, at least) in `development` and I'd consider this a road-block for 0.3.0. So either we fix...