Add test

test/core/cudadrv.jl

@@ -200,6 +200,34 @@ end
 @test (CUDA.@elapsed identity(nothing)) > 0
 @test (CUDA.@elapsed blocking=true identity(nothing)) > 0
 
+let
+    function simple_matmul(C, A, B, N)
+        for i = 1:N
+            for j = 1:N
+                elem = C[i, j]
+
+                for k = 1:N
+                    elem += A[i, k] * B[k, j]
+                end
+
+                C[i, j] = elem
+            end
+        end
+
+        nothing
+    end
+
+    A = CuArray(rand(Float32, (1000, 1000)))
+    B = CuArray(rand(Float32, (1000, 1000)))
+    C = similar(A)
+
+    N1 = (CUDA.@elapsed @cuda simple_matmul(C, A, B, 500))
+    N2 = (CUDA.@elapsed @cuda simple_matmul(C, A, B, 1000))
+
+    @test 16 * N1 > N2
+    @test N2 > 4 * N1
+end
+
 CuEvent(CUDA.EVENT_BLOCKING_SYNC)
 CuEvent(CUDA.EVENT_BLOCKING_SYNC | CUDA.EVENT_DISABLE_TIMING)