rostime call is slow

While developing a Matlab wrapper around a ROS package, I noticed that rostime("now") takes around 1.4e-03 s per call. In Python on the other hand the same call only takes 4.8e-06 s. I understand that there might be implementation differences, but Matlab being orders of magnitude slower than Python seems strange. Is there anything I am doing wrong here?

Matlab script for timing:

tic
for i = 1:1000
    a = rostime("now");
end
toc / 1000

Python script for timing:

import rospy
import time

if __name__ == "__main__":
    rospy.init_node("test")

    t0 = time.clock()
    for i in range(1000):
        a = rospy.Time.now()
    t1 = time.clock()
    print((t1 - t0) / 1000)

Thanks!

Edit:

I found this bug report and the workaround using robotics.ros.internal.Time([]).CurrentTime is about 10 times faster than rostime("now"). This is better but still not as fast as it should be. I am also not sure how much of a hack this is.

Answer 1

[Amrtanshu Raj]

Hi,

The difference in computation time is because of the implementation difference between the MATLAB and Python version of the command. Even declaring a static rostime like

time = rostime(2,85000); 

Takes around 1.4e-03 seconds to compute. I have brought this to the notice of the developers.

Thanks !!