This reduces logging overhead. The performance difference is only significant when there are many fast nodes. In my test file with many math nodes, the performance improved from 720ms to 630ms.
6.9 KiB
6.9 KiB