diff --git a/test/examples/gpu.jl b/test/examples/gpu.jl
index c99c67f13..7aca5f83d 100644
--- a/test/examples/gpu.jl
+++ b/test/examples/gpu.jl
@@ -136,14 +136,15 @@ end
                                                parallelization_backend=Main.parallelization_backend)
 
             # Note that this simulation only takes 42 time steps on the CPU.
-            # Due to https://github.com/JuliaGPU/Metal.jl/issues/549, it doesn't work on Metal.
+            # TODO This takes 43 time steps on Metal.
+            # Maybe related to https://github.com/JuliaGPU/Metal.jl/issues/549
             trixi_include_changeprecision(Float32, @__MODULE__,
                                           joinpath(examples_dir(), "fluid",
                                                    "dam_break_3d.jl"),
                                           tspan=(0.0f0, 0.1f0),
                                           fluid_particle_spacing=0.1,
                                           semi=semi_fullgrid,
-                                          maxiters=42)
+                                          maxiters=43)
             @test sol.retcode == ReturnCode.Success
             backend = TrixiParticles.KernelAbstractions.get_backend(sol.u[end].x[1])
             @test backend == Main.parallelization_backend