Aug 6, 2019, 8:05 PM by [email protected]: > Hi! > > Sorry about beeing unclear. I know that I switched the underlaying RT os. > > The jessie box runs: > Linux beaglebone 3.8.13-xenomai-r83 #1 Thu Jan 5 01:32:32 UTC 2017 armv7l > GNU/Linux > > And on stretch: > Linux beaglebone 4.19.37-bone-rt-r33 > #1stretch <>> PREEMPT RT Wed Jun 5 > 11:22:07 UTC 2019 armv7l GNU/Linux > > I switched all driver/config stuff to use the preempt rt stuff. > > The jessie box is perfectly fine. The stretch based one is unusable. > I am wondering how others use a stretch based image and if I am doing > something wrong > and there is a "magic" kernel version that is fast enough to be used. > > I did a simple benchmark opening a text engraving gcode file from fusion360. > On the Jessie box it takes ~20s from clicking load to display. > On the stretch base box takes around 4 minutes... > > Or could it have something to do with the fact that the stretch image runs > from SD and the jessie from internal flash? > Is there so much disk I/O involved? > Well, it pretty well could be. To reach some verdict, the inputs should be the same. Difference of 3.40 minutes seems little too excessive for just change in kernel and distribution versions. However, you can test it all (if you have the time and incentive). You can install the 3.8 Linux Kernel with 2.x Xenomai patch on Debian Stretch and vice versa. (But you will - probably - have to do the compiling yourself.)
Cern. -- website: http://www.machinekit.io blog: http://blog.machinekit.io github: https://github.com/machinekit --- You received this message because you are subscribed to the Google Groups "Machinekit" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/machinekit/LlgjPzb--3-1%40tuta.io.
