fix

Author: CarloLucibello

GNNlib/test/msgpass.jl

@@ -168,8 +168,7 @@ end
 
     @testset "copy_xj +" begin
         for g in TEST_GRAPHS
-            dev = gpu_device(force=true)
-            broken = get_graph_type(g) == :sparse && dev isa AMDGPUDevice
+            broken = get_graph_type(g) == :sparse && gpu_backend() == "AMDGPU"
             f(g, x) = propagate(copy_xj, g, +, xj = x)
             @test test_gradients(
                 f, g, g.x; test_gpu=true, test_grad_f=false, compare_finite_diff=false
@@ -179,8 +178,7 @@ end
 
     @testset "copy_xj mean" begin
         for g in TEST_GRAPHS
-            dev = gpu_device(force=true)
-            broken = get_graph_type(g) == :sparse && dev isa AMDGPUDevice
+            broken = get_graph_type(g) == :sparse && gpu_backend() == "AMDGPU"
             f(g, x) = propagate(copy_xj, g, mean, xj = x)
             @test test_gradients(
                 f, g, g.x; test_gpu=true, test_grad_f=false, compare_finite_diff=false
@@ -190,8 +188,7 @@ end
 
     @testset "e_mul_xj +" begin
         for g in TEST_GRAPHS
-            dev = gpu_device(force=true)
-            broken = get_graph_type(g) == :sparse && dev isa AMDGPUDevice
+            broken = get_graph_type(g) == :sparse && gpu_backend() == "AMDGPU"
             e = rand(Float32, size(g.x, 1), g.num_edges)
             f(g, x, e) = propagate(e_mul_xj, g, +; xj = x, e)
             @test test_gradients(
@@ -207,7 +204,6 @@ end
                 g = set_edge_weight(g, w)
                 return propagate(w_mul_xj, g, +, xj = x)
             end
-            dev = gpu_device(force=true)
             # @show get_graph_type(g) has_isolated_nodes(g)
             # broken = get_graph_type(g) == :sparse
             broken = true

GNNlib/test/test_module.jl

@@ -45,7 +45,7 @@ using Flux: Flux
 # from this module
 export D_IN, D_OUT, GRAPH_TYPES, TEST_GRAPHS,
        test_gradients, finitediff_withgradient, 
-       check_equal_leaves
+       check_equal_leaves, gpu_backend
 
 
 const D_IN = 3
@@ -177,4 +177,18 @@ TEST_GRAPHS = [generate_test_graphs(:coo)...,
                generate_test_graphs(:dense)...,
                generate_test_graphs(:sparse)...]
 
+
+function gpu_backend()
+    dev = gpu_device()
+    if dev isa CUDADevice
+        return "CUDA"
+    elseif dev isa AMDGPUDevice
+        return "AMDGPU"
+    elseif dev isa MetalDevice
+        return "Metal"
+    else
+        return "Unknown"
+    end
+end
+
 end # module

GraphNeuralNetworks/test/layers/conv.jl

@@ -102,7 +102,7 @@ end
     l = ChebConv(D_IN => D_OUT, k)
     for g in TEST_GRAPHS
         has_isolated_nodes(g) && continue
-        broken = get_graph_type(g) == :sparse || gpu_device() isa AMDGPUDevice
+        broken = get_graph_type(g) == :sparse || gpu_backend() == "AMDGPU"
         @test size(l(g, g.x)) == (D_OUT, g.num_nodes) broken=broken
         @test test_gradients(
             l, g, g.x, rtol = RTOL_LOW, test_gpu = true, compare_finite_diff = false

GraphNeuralNetworks/test/test_module.jl

@@ -38,9 +38,15 @@ using SparseArrays
 # from Base
 export mean, randn, SparseArrays, AbstractSparseMatrix
 
-# from other packages
-export Flux, gradient, Dense, Chain, relu, random_regular_graph, erdos_renyi,
-       BatchNorm, LayerNorm, Dropout, Parallel
+# from Flux.jl
+export Flux, gradient, Dense, Chain, relu
+       BatchNorm, LayerNorm, Dropout, Parallel,
+       gpu_device, cpu_device, get_device,
+       CPUDevice, CUDADevice, AMDGPUDevice, MetalDevice,
+       gpu_backend
+
+# from Graphs.jl
+export random_regular_graph, erdos_renyi
 
 # from this module
 export D_IN, D_OUT, GRAPH_TYPES, TEST_GRAPHS,
@@ -178,5 +184,19 @@ TEST_GRAPHS = [generate_test_graphs(:coo)...,
                generate_test_graphs(:dense)...,
                generate_test_graphs(:sparse)...]
 
+
+function gpu_backend()
+    dev = gpu_device()
+    if dev isa CUDADevice
+        return "CUDA"
+    elseif dev isa AMDGPUDevice
+        return "AMDGPU"
+    elseif dev isa MetalDevice
+        return "Metal"
+    else
+        return "Unknown"
+    end
+end
+
 end # testmodule