Merge pull request #114 from JordiManyer/bugfix-global-length-prange

Bugfix: `global_length` for `PRange`

Author: fverdugo

CHANGELOG.md

@@ -11,6 +11,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 - Added an MPI ibarrier-based (supposedly scalable) algorithm to find rcv neighbours in a sparse all-to-all communication graph given the snd neighbors. We left the previous non-scalable algorithm as default (based on gather-scatter) until we have experimental evidence on the relative performance and scalability of the former with respect to the latter and for which core ranges.
 
+### Fixed
+
+- Bugfix: `global_length` for `PRange` not working as intended. 
+
 ## [0.3.2] - 2023-05-10
 
 ### Fixed

src/p_range.jl

@@ -1642,7 +1642,7 @@ end
 
 ##prange(f,args...) = PRange(f(args...))
 
-global_length(pr::PRange) = map(local_length,partition(pr))
+global_length(pr::PRange) = map(global_length,partition(pr))
 local_length(pr::PRange) = map(local_length,partition(pr))
 own_length(pr::PRange) = map(own_length,partition(pr))
 local_to_global(pr::PRange) = map(local_to_global,partition(pr))

test/p_range_tests.jl

@@ -190,6 +190,7 @@ function p_range_tests(distribute)
 
    ids4 = uniform_partition(parts,(2,2),(5,4))
    @test length(PRange(ids4)) == 4*5
+   @test PartitionedArrays.getany(global_length(PRange(ids4))) == 4*5
    map(parts,ids4) do part, lid_to_gid
        if part == 1
            @test lid_to_gid == [1, 2, 6, 7]