EventWorkspace: iterator to go through all events (?)

[Janik Zikovsky]

Will allow operations (weights, etc.) to be applied to all events in a single loop.

A forward iterator would be the quickest ( O(1) to go forward ). But it is not possible to parallelize a forward iterator with OpenMP.

If made a random-access iterator, then you could do a OpenMP for loop; but unfortunately each access will have complexity ( O(log M) ) where M is the number of spectra.

Therefore, a mix of both methods might be necessary - split into X ForwardIterators that operate in parallel, where X is the number of available cores?

[Nick Draper]

We found with histogram workspaces that iterators were generally slower than nested loops, even when optimised we could not beat the speed of a nested loop.

Also as you said, for loops work well with openMP

[Stuart Campbell]

This ticket has been transferred to github issue ​3031