Page 1 of 1

ccr1009 speedtest resoults

Posted: Fri Dec 14, 2018 5:27 pm
by zakynthoswifi
Hi,
i'm running a speedtest on loopback address and the resoult is this
ccr1009 speedtest.png
Is this resoult correct?
also if i run a speedtest from a rb3011 with fiber uplink connected directly to ccr over 1G sfp i'm getting 459,32 Mbps max (one direction) if i setect both directions then i get 30Mbps Rx and 7,4 Mbps Tx

Re: ccr1009 speedtest resoults

Posted: Sat Dec 15, 2018 2:32 am
by mducharme
The btest process itself uses quite a bit of CPU. Results will always be lower than the device supports if the CPU is being maxed out or nearly maxed out by the btest process.

Re: ccr1009 speedtest resoults

Posted: Sat Dec 15, 2018 4:24 pm
by trace323
The btest process itself uses quite a bit of CPU. Results will always be lower than the device supports if the CPU is being maxed out or nearly maxed out by the btest process.
This is false. the CCR only goes to 1%.

I notice they are issues with the CCR with speed now days. I have a 500/500 Fiber Pipe

I only get 500MbpsDown/200 UP at the same time. Separately I get the correct speeds.

I am conducting a very long test and survey because I know there is some sort of firmware issue now or hardware issue now.

I know for a fact there is a problem now.

Re: ccr1009 speedtest resoults

Posted: Sun Dec 16, 2018 5:13 am
by mducharme
The btest process itself uses quite a bit of CPU. Results will always be lower than the device supports if the CPU is being maxed out or nearly maxed out by the btest process.
This is false. the CCR only goes to 1%.
That is not false. The btest server only became multithreaded as of 6.44beta39. Before this version it would only run on one core, so if you max out even one core of a CCR router then the bandwidth test will be limited by that. Did you run tool profile to see what your individual CPU usage was when running the btest? CCR routers have many weak cores instead of a few powerful ones. On a CCR1072 you could have one core maxed out and the other 71 cores with low usage giving you only a few percent total CPU usage. With CCR you should ignore the total CPU usage and look at usage of the individual cores instead.