HM2_ETH >> iptables >> hm2−eth−rules−output
01 Mar 2021 14:10 #200692
by PCW
Replied by PCW on topic HM2_ETH >> iptables >> hm2−eth−rules−output
Note that disabling RT CPUs with isolcpus may in fact make your
Ethernet and LinuxCNC latency worse.
The latency test really only tests dispatch latency, overall LinuxCNC latency
actually doing things will typically be significantly worse. Watching the
.tmax values shows the actual latency
Ethernet and LinuxCNC latency worse.
The latency test really only tests dispatch latency, overall LinuxCNC latency
actually doing things will typically be significantly worse. Watching the
.tmax values shows the actual latency
The following user(s) said Thank You: robertspark
Please Log in or Create an account to join the conversation.
- robertspark
- Offline
- Platinum Member
-
Less
More
- Posts: 915
- Thank you received: 216
01 Mar 2021 15:43 #200705
by robertspark
Thanks.... more investigation required. (seem to have got rid of the "hm2/hm2_7176e.0: error finishing red! iter=6" once ethtool -C worked and the ping times dropped from 1.02msec to 0.200mSec).
I'll trythis evening
All part of the learning process for me and I'm happy to learn.
{I've not tried it, but I was wondering if intel_pstate=disable.... etc mce=ignore_ce make any difference}
Replied by robertspark on topic HM2_ETH >> iptables >> hm2−eth−rules−output
Note that disabling RT CPUs with isolcpus may in fact make your
Ethernet and LinuxCNC latency worse.
The latency test really only tests dispatch latency, overall LinuxCNC latency
actually doing things will typically be significantly worse. Watching the
.tmax values shows the actual latency
Thanks.... more investigation required. (seem to have got rid of the "hm2/hm2_7176e.0: error finishing red! iter=6" once ethtool -C worked and the ping times dropped from 1.02msec to 0.200mSec).
I'll try
halcmd show param *.tmax
All part of the learning process for me and I'm happy to learn.
{I've not tried it, but I was wondering if intel_pstate=disable.... etc mce=ignore_ce make any difference}
Please Log in or Create an account to join the conversation.
Moderators: cmorley
Time to create page: 0.067 seconds