this is not a test... (without substantiation) run;
speedtest-ookla
if you want to test (internet) speed from the router
thanks.
got 800+ mbps on the router, 450+ on wired-connected client.
so i think i'm there.
p.s thanks for your build!
1 Like
you deleted 6 posts of yours from this thread...
doing so alters the context of many replies...
-
'speedtest-ookla' is specific to my build which the OP stated they had installed (removed)
-
deletion of posts is typically inbounds for privacy reasons(editing is better if done within the first 2mins) or lack of relevance to anything else in the thread (or just plain mistake) in your case... perhaps you saw them as no longer being relevant... this is not the case...
-
people reading this thread in the future need to be able to see all context with it's original intent... as even if a line of discussion proved not so useful to you in this instance... it may to others for their similar situations... ( you can edit a post with the phrase 'edit: this turned out to be not relevant' or similar if you feel future readers may not benefit from it... or make use of the gear symbol to select an area and choose 'hide details' to de-emphasise )
-
users of this forum dedicated their personal time and energy giving you advise and responding to your posts... deleting them means that in the future they are less likely to help you or others...
5 Likes
If I had to hazard a guess now, the following was true:
- User did not have irqbalance properly installed/enabled, thus all interrupts were pinned to CPU 0 (default)
- User wasn't testing with iperf using two distinct endpoints, but was instead using the router as one of the endpoints, which causes the Heisenberg effect...iperf itself will run on CPU 0, thus taking away resources from the router to actually do its job of shuttling packets and thus negatively impact testing
This topic was automatically closed 10 days after the last reply. New replies are no longer allowed.