use only 256 and 512 nodes for convergence test

Author: JoshuaLampert

test/test_unit.jl

@@ -540,15 +540,15 @@ end
 
     accuracy_orders = [2, 4, 6]
     for accuracy_order in accuracy_orders
-        eoc_mean_values, _ = convergence_test(@__MODULE__, default_example(), 2, N = 512,
+        eoc_mean_values, _ = convergence_test(@__MODULE__, default_example(), 2, N = 256,
                                               tspan = (0.0, 1.0),
                                               accuracy_order = accuracy_order)
         @test isapprox(eoc_mean_values[:l2][1], accuracy_order, atol = 0.1)
         @test isapprox(eoc_mean_values[:linf][1], accuracy_order, atol = 0.1)
         @test isapprox(eoc_mean_values[:l2][2], accuracy_order, atol = 0.1)
         @test isapprox(eoc_mean_values[:linf][2], accuracy_order, atol = 0.1)
 
-        eoc_mean_values2, _ = convergence_test(@__MODULE__, default_example(), [512, 1024],
+        eoc_mean_values2, _ = convergence_test(@__MODULE__, default_example(), [256, 512],
                                                tspan = (0.0, 1.0),
                                                accuracy_order = accuracy_order)
         for kind in (:l2, :linf), variable in (1, 2)