Results of latency test, list of computers tested for use with LinuxCNC

boozeman
good looking! Would be suitable even for parport, I guess.
I have a set of similar hp elitedesk 800 G2, also sporting an i5 6500. Will test them soon
greez
chris

7i76S ???
To be sure it will work properly for longer, since you already have the 7i96S i assume, wire the +5V and ethernet cable, run linuxcnc and leave it on for a day or two while occasionally doing something else on the PC.
I usually leave it for 2 to 5 days just to be sure.

  Note: Using LXRT realtime

MB  GA-J1800N-D2H(BIOS F6)
RAM KVR13S9S8/4 4Gb
SSD SA400S37/240G
Debian Wheezy 3.2.0-6-rt-686-pae
LinuxCNC v. 2.8.4 rip build

kernel parameters: "isolcpus=0 intel_idle.max_cstate=0 processor.max_cstate=0 idle=poll quiet"
 
kernel parameters: "isolcpus=1 intel_idle.max_cstate=0 processor.max_cstate=0 idle=poll quiet"
 
 

Old test of this hardware(with HDD).
Debian Wheezy 3.2.0-6-rt-686-pae
LinuxCNC v. 2.7.14

kernel parameters: "isolcpus=0 lapic quiet"
 

kernel parameters: "isolcpus=1 lapic quiet"
 

D.L. wrote:

Old test of this hardware(with HDD).
Debian Wheezy 3.2.0-6-rt-686-pae
LinuxCNC v. 2.7.14

kernel parameters: "isolcpus=0 lapic quiet"

kernel parameters: "isolcpus=1 lapic quiet"
 
 

D.L.
The results are nice and seem to support my assumption, rotational discs hamper latency - on first sight. A closer look shows, one test is with 5 glxgears and the other with 10. So: are these two results really comparable in terms of reproducibility?

I am working on updating the script as promised right now. No ETA so far. It will address the bugs recorded up to date and have some enhancements too. So, stay tuned
greez chris