try new test approach for benchmark

test/runtests.jl

@@ -38,4 +38,54 @@ include("transform.jl")
 include("utils.jl")
 
 # Benchmark the transform function
-include("benchmarks.jl")
+@testset "Benchmarks" begin
+    s = 10 # number of samples
+    e = 1 # number of evals
+
+    # Create an array of input sizes to benchmark for 2D
+    sizes_2D = [2^i for i in 3:4]
+
+    # Create DataFrames to store the benchmark results
+    df_dt_2D = DataFrame(
+        os_info = os_info,
+        gpu_info = gpu_info,
+        sizes = sizes_2D,
+        dt_proposed = Float64[]
+    )
+
+    for n in sizes_2D
+        f = Float32.(rand([0, 1], n, n))
+
+        if dev != Array
+            f_dev = dev(f)
+            dt = @benchmark(transform($boolean_indicator($f_dev)); samples=s, evals=e)
+            push!(df_dt_2D, [os_info, gpu_info, n, minimum(dt).time])
+        end
+    end
+
+    # Create an array of input sizes to benchmark for 3D
+    sizes_3D = [2^i for i in 0:2]
+
+    # Create DataFrames to store the benchmark results
+    df_dt_3D = DataFrame(
+        os_info = os_info,
+        gpu_info = gpu_info,
+        sizes = sizes_3D,
+        dt_proposed = Float64[]
+    )
+
+    for n in sizes_3D
+        f = Float32.(rand([0, 1], n, n, n))
+
+        if dev != Array
+            f_dev = dev(f)
+            dt = @benchmark(transform($boolean_indicator($f_dev)); samples=s, evals=e)
+            push!(df_dt_3D, [os_info, gpu_info, n, minimum(dt).time])
+        end
+    end
+
+    # Show the dataframes
+    @info df_dt_2D
+    @info df_dt_3D
+    @test 1 == 1
+end