On Sunday, 7 August 2016 at 22:43:35 UTC, Isaac Gouy wrote:
"Chapel programs versus C++ g++" just compares 2. Where are those "more specific benchmarks" ?
I think he's referring to a HPC cluster running chapel vs. C++ versions of programs.
However, I would argue Chapel would only win those comparisons when you're comparing Chapel taking advantage of all the different computers in the cluster vs. C++ code that is not taking advantage of everything it could. For instance, Chapel does not have good support for graphics cards as of yet. Anyway, Chapel does their own benchmarks against things like C+MPI. In some cases it's good, but it's not consistently better. The advantage to Chapel is its simplicity and expressiveness.
