Loop Timer
Measuring loop speed without slowing the loop
Faster loops mean tighter path following. The loop timer shows how fast your OpMode really runs, at a cost of one System.nanoTime() and a few additions per loop (about 50 ns). It never allocates and never sends telemetry; you decide what to print and when.
Use
A follower built with dromos.localizer(...) already ticks a timer inside every follower.update():
import com.pedropathing.follower.Follower;
import me.munchem.electrodromos.LoopTimer;
import me.munchem.electrodromos.Dromos;
public class MyTeleOp extends OpMode {
private Follower follower;
private LoopTimer loop;
@Override
public void init() {
follower = Constants.create(hardwareMap); // Pedro follower, built with ElectroDromos
loop = Dromos.loopTimer(follower);
}
@Override
public void start() {
loop.reset(); // don't count the wait between init and start
}
@Override
public void loop() {
follower.manual(-gamepad1.left_stick_y, gamepad1.left_stick_x, gamepad1.right_stick_x);
follower.update(); // ticks the timer
telemetry.addData("loop Hz", loop.hz());
telemetry.addData("worst loop (ms)", loop.maxMs());
telemetry.update();
}
}Dromos.loopTimer(follower) throws if the follower's localizer didn't come from dromos.localizer(...).
Readings
Stats are published once per 500 ms window, so they stay still long enough to read. They're 0 until the first window completes.
| Method | Meaning |
|---|---|
hz() | Loops per second over the last window |
averageMs() | Average loop time over the last window |
maxMs() | Slowest loop in the last window: the spikes that hurt path following |
lastMs() | The most recent loop |
reset() | Forget everything, e.g. in start() |
Telemetry is the expensive part
Measuring costs nothing noticeable. Sending telemetry every loop can cost milliseconds. If the loop timer shows telemetry slowing you down, send it less often. You'll see the loop speed change in the timer itself.
Timing a loop without a follower
LoopTimer loop = new LoopTimer();
// once per loop:
loop.tick();Only tick once per loop. If you also call follower.update(), use Dromos.loopTimer(follower) instead of a second timer.
Last updated on