Feature support testing for IAllreduce and IReduce for GPU backends

Author: Keluaa

test/mpi_support_test.jl

@@ -0,0 +1,28 @@
+include("common.jl")
+
+MPI.Init()
+
+# Those MPI calls may be unsupported features (e.g. for GPU backends), and will raise SIGSEGV
+# (or a similar signal) when called, which cannot be handled in Julia in a portable way.
+
+op = ARGS[1]
+if op == "IAllreduce"
+    # IAllreduce is unsupported for CUDA with OpenMPI + UCX
+    # See https://docs.open-mpi.org/en/main/tuning-apps/networking/cuda.html#which-mpi-apis-do-not-work-with-cuda-aware-ucx
+    send_arr = ArrayType(zeros(Int, 1))
+    recv_arr = ArrayType{Int}(undef, 1)
+    synchronize()
+    req = MPI.IAllreduce!(send_arr, recv_arr, +, MPI.COMM_WORLD)
+    MPI.Wait(req)
+
+elseif op == "IReduce"
+    # IAllreduce is unsupported for CUDA with OpenMPI + UCX
+    send_arr = ArrayType(zeros(Int, 1))
+    recv_arr = ArrayType{Int}(undef, 1)
+    synchronize()
+    req = MPI.IReduce!(send_arr, recv_arr, +, MPI.COMM_WORLD; root=0)
+    MPI.Wait(req)
+
+else
+    error("unknown test: $op")
+end

test/runtests.jl

@@ -71,6 +71,19 @@ if Sys.isunix()
     include("mpiexecjl.jl")
 end
 
+function is_mpi_operation_supported(mpi_op, n=nprocs)
+    test_file = joinpath(@__DIR__, "mpi_support_test.jl")
+    cmd = `$(mpiexec()) -n $n $(Base.julia_cmd()) --startup-file=no $test_file $mpi_op`
+    supported = success(run(ignorestatus(cmd)))
+    !supported && @warn "$mpi_op is unsupported with $backend_name"
+    return supported
+end
+
+if ArrayType != Array  # we expect that only GPU backends can have unsupported features
+    ENV["JULIA_MPI_TEST_IALLREDUCE"] = is_mpi_operation_supported("IAllreduce")
+    ENV["JULIA_MPI_TEST_IREDUCE"] = is_mpi_operation_supported("IReduce")
+end
+
 excludefiles = split(get(ENV,"JULIA_MPI_TEST_EXCLUDE",""),',')
 
 testdir = @__DIR__

test/test_allreduce.jl

@@ -13,6 +13,9 @@ else
     operators = [MPI.SUM, +, (x,y) -> 2x+y-x]
 end
 
+iallreduce_supported = get(ENV, "JULIA_MPI_TEST_IALLREDUCE", "true") == "true"
+
+
 for T = [Int]
     for dims = [1, 2, 3]
         send_arr = ArrayType(zeros(T, Tuple(3 for i in 1:dims)))
@@ -46,16 +49,20 @@ for T = [Int]
 
             # Nonblocking
             recv_arr = ArrayType{T}(undef, size(send_arr))
-            req = MPI.IAllreduce!(send_arr, recv_arr, op, MPI.COMM_WORLD)
-            MPI.Wait(req)
-            @test recv_arr == comm_size .* send_arr
+            if iallreduce_supported
+                req = MPI.IAllreduce!(send_arr, recv_arr, op, MPI.COMM_WORLD)
+                MPI.Wait(req)
+            end
+            @test recv_arr == comm_size .* send_arr skip=!iallreduce_supported
 
             # Nonblocking (IN_PLACE)
             recv_arr = copy(send_arr)
             synchronize()
-            req = MPI.IAllreduce!(recv_arr, op, MPI.COMM_WORLD)
-            MPI.Wait(req)
-            @test recv_arr == comm_size .* send_arr
+            if iallreduce_supported
+                req = MPI.IAllreduce!(recv_arr, op, MPI.COMM_WORLD)
+                MPI.Wait(req)
+            end
+            @test recv_arr == comm_size .* send_arr skip=!iallreduce_supported
         end
     end
 end

test/test_reduce.jl

@@ -9,6 +9,8 @@ const can_do_closures =
     Sys.ARCH !== :aarch64 &&
     !startswith(string(Sys.ARCH), "arm")
 
+ireduce_supported = get(ENV, "JULIA_MPI_TEST_IREDUCE", "true") == "true"
+
 using DoubleFloats
 
 MPI.Init()
@@ -119,18 +121,22 @@ for T = [Int]
 
             # Nonblocking
             recv_arr = ArrayType{T}(undef, size(send_arr))
-            req = MPI.IReduce!(send_arr, recv_arr, op, MPI.COMM_WORLD; root=root)
-            MPI.Wait(req)
+            if ireduce_supported
+                req = MPI.IReduce!(send_arr, recv_arr, op, MPI.COMM_WORLD; root=root)
+                MPI.Wait(req)
+            end
             if isroot
-                @test recv_arr == sz .* send_arr
+                @test recv_arr == sz .* send_arr skip=!ireduce_supported
             end
 
             # Nonblocking (IN_PLACE)
             recv_arr = copy(send_arr)
-            req = MPI.IReduce!(recv_arr, op, MPI.COMM_WORLD; root=root)
-            MPI.Wait(req)
+            if ireduce_supported
+                req = MPI.IReduce!(recv_arr, op, MPI.COMM_WORLD; root=root)
+                MPI.Wait(req)
+            end
             if isroot
-                @test recv_arr == sz .* send_arr
+                @test recv_arr == sz .* send_arr skip=!ireduce_supported
             end
         end
     end
@@ -148,10 +154,12 @@ else
 end
 
 recv_arr = isroot ? zeros(eltype(send_arr), size(send_arr)) : nothing
-req = MPI.IReduce!(send_arr, recv_arr, +, MPI.COMM_WORLD; root=root)
-MPI.Wait(req)
+if ireduce_supported
+    req = MPI.IReduce!(send_arr, recv_arr, +, MPI.COMM_WORLD; root=root)
+    MPI.Wait(req)
+end
 if rank == root
-    @test recv_arr ≈ [Double64(sz*i)/10 for i = 1:10] rtol=sz*eps(Double64)
+    @test recv_arr ≈ [Double64(sz*i)/10 for i = 1:10] rtol=sz*eps(Double64) skip=!ireduce_supported
 end
 
 MPI.Barrier( MPI.COMM_WORLD )