What can cause the Ping time to my server (running on the same machine as the client) to increase to a large value? For a period of time, it's 0-1-2 sec to ping my server, then "something happens" and it increases to a high value--17, 18, 20 ms with an overall delay beyond usability. I have written a bug report because I am quite sure that minimizing the Jamulus Server window can do this, but there's something else as well, which I haven't been able to chase down.
What could it be?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Other programs running on the PC? background services running?
Are you addressing the server from the client as localhost (or 127.0.0.1)?
If not on localhost are you using a wifi connection to the net?
My next step would be to be certain that the server is addressed locally, if that doesn't help kill any other programs running other the the Jamulus server and client.
I have not seen such behaviour when running server and client on the same compi.
I have, however exact that behaviour when accessing an external server with a wifi connection, even a very good and fast wifi connection. Using a cable clears it up.
I hope this helps,
Don
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Lots of CPU left, according to Activity monitor. Chrome running, minimized.
I tried both addressing the server as localhost and not, and there was
little difference, perhaps localhost slightly slower. Wired connection, not
WiFi. Wired mic.
Other programs running on the PC? background services running?
Are you addressing the server from the client as localhost (or 127.0.0.1)?
If not on localhost are you using a wifi connection to the net?
My next step would be to be certain that the server is addressed locally,
if that doesn't help kill any other programs running other the the Jamulus
server and client.
I have not seen such behaviour when running server and client on the same
compi.
I have, however exact that behaviour when accessing an external server
with a wifi connection, even a very good and fast wifi connection. Using a
cable clears it up.
I hope this helps,
Don
Do we know the reaction time of the activity monitor? I would not trust it to find possibly very short timing problems or artifacts in poorly designed programs that could be holding things up (who knows). For testing anyway I would turn off everything possible and then reboot just to see if it makes a difference.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
The Activity Monitor takes a lot of CPU cycles. I know not to rely on it. I checked it just to see whether there was something running that I wasn't aware of that was taking a lot of the CPU time, but there was nothing other than Jamulus. The effect I am seeing is not short-lived, but characterized by a gradual increase over minutes, and then a decrease initiated by (maybe) opening & closing the Settings window??
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Sorry, but I don't have any other good ideas.
The only thing I see by me is every time I change to buffer settings the delay goes up a bit and then goes down again over maybe 10 seconds. The ping times depend, of course, on the server I am connected to. If I connect to a server on this compi with localhost the ping is 0 (as I would hope, there is no network traffic) and the total delay also starts at about 18ms and then descends to 15ms and stays there.
You've got another of those mysteries of modern technology.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Sorry, but I don't have any other good ideas.
The only thing I see by me is every time I change to buffer settings the
delay goes up a bit and then goes down again over maybe 10 seconds. The
ping times depend, of course, on the server I am connected to. If I connect
to a server on this compi with localhost the ping is 0 (as I would hope,
there is no network traffic) and the total delay also starts at about 18ms
and then descends to 15ms and stays there.
You've got another of those mysteries of modern technology.
What can cause the Ping time to my server (running on the same machine as the client) to increase to a large value? For a period of time, it's 0-1-2 sec to ping my server, then "something happens" and it increases to a high value--17, 18, 20 ms with an overall delay beyond usability. I have written a bug report because I am quite sure that minimizing the Jamulus Server window can do this, but there's something else as well, which I haven't been able to chase down.
What could it be?
What hardware are you using? What OS?
13” MacBook late 2015, MacOS 10.15.4.
On Sat, May 23, 2020 at 02:12 Luuk luuk@users.sourceforge.net wrote:
Other programs running on the PC? background services running?
Are you addressing the server from the client as localhost (or 127.0.0.1)?
If not on localhost are you using a wifi connection to the net?
My next step would be to be certain that the server is addressed locally, if that doesn't help kill any other programs running other the the Jamulus server and client.
I have not seen such behaviour when running server and client on the same compi.
I have, however exact that behaviour when accessing an external server with a wifi connection, even a very good and fast wifi connection. Using a cable clears it up.
I hope this helps,
Don
Lots of CPU left, according to Activity monitor. Chrome running, minimized.
I tried both addressing the server as localhost and not, and there was
little difference, perhaps localhost slightly slower. Wired connection, not
WiFi. Wired mic.
I will try quitting Chrome and see if it matters.
On Sat, May 23, 2020 at 09:06 DonC doncor@users.sourceforge.net wrote:
Do we know the reaction time of the activity monitor? I would not trust it to find possibly very short timing problems or artifacts in poorly designed programs that could be holding things up (who knows). For testing anyway I would turn off everything possible and then reboot just to see if it makes a difference.
The Activity Monitor takes a lot of CPU cycles. I know not to rely on it. I checked it just to see whether there was something running that I wasn't aware of that was taking a lot of the CPU time, but there was nothing other than Jamulus. The effect I am seeing is not short-lived, but characterized by a gradual increase over minutes, and then a decrease initiated by (maybe) opening & closing the Settings window??
Sorry, but I don't have any other good ideas.
The only thing I see by me is every time I change to buffer settings the delay goes up a bit and then goes down again over maybe 10 seconds. The ping times depend, of course, on the server I am connected to. If I connect to a server on this compi with localhost the ping is 0 (as I would hope, there is no network traffic) and the total delay also starts at about 18ms and then descends to 15ms and stays there.
You've got another of those mysteries of modern technology.
I think what I'm seeing is a bug, but I don't know what might be causing
it. Maybe someone else will have experienced it also...?
On Sat, May 23, 2020 at 12:50 PM DonC doncor@users.sourceforge.net wrote: