Hi,
After calling pjsua_set_snd_dev() CPU loads up to 10%. Mac OS X 10.5.5, Intel.
Is that an ussue or a normal behavior?
--
Best regards,
Alexei Kuznetsov
Hi,
I guess that's normal, I experienced about the same with echo
canceller enabled (running pjsua without any params).
Regards,
nanang
On Thu, Nov 6, 2008 at 5:45 AM, Alexei Kuznetsov eofster@gmail.com wrote:
Hi,
After calling pjsua_set_snd_dev() CPU loads up to 10%. Mac OS X 10.5.5, Intel.
Is that an ussue or a normal behavior?
--
Best regards,
Alexei Kuznetsov
Visit our blog: http://blog.pjsip.org
pjsip mailing list
pjsip@lists.pjsip.org
http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org
I mean, the application uses 10% of cpu without any media active.
Also, pjsua without any params uses less than 1% cpu after start-up,
but 10% after the first call. Is it something you could improve in the
future? Or is it the design, or is it not up to pjproject but to other
code pjrpoject uses? Is it on a Mac only?
--
Best regards,
Alexei
On Thu, Nov 6, 2008 at 12:05 PM, Nanang Izzuddin nanang@pjsip.org wrote:
Hi,
I guess that's normal, I experienced about the same with echo
canceller enabled (running pjsua without any params).
Regards,
nanang
On Thu, Nov 6, 2008 at 5:45 AM, Alexei Kuznetsov eofster@gmail.com wrote:
Hi,
After calling pjsua_set_snd_dev() CPU loads up to 10%. Mac OS X 10.5.5, Intel.
Is that an ussue or a normal behavior?
--
Best regards,
Alexei Kuznetsov
Hi Alexei,
On 06 Nov 2008, at 13:56, Alexei Kuznetsov wrote:
I mean, the application uses 10% of cpu without any media active.
Also, pjsua without any params uses less than 1% cpu after start-up,
but 10% after the first call. Is it something you could improve in the
future? Or is it the design, or is it not up to pjproject but to other
code pjrpoject uses? Is it on a Mac only?
--
Best regards,
Alexei
I've noticed exactly the same behaviour on Mac, it's a very annoying
problem. Just having the sound device open seems to consume an
inordinate amount of CPU cycles already. From what I've seen, it
actually seems to be a PortAudio issue. I already went on their
mailinglist to discuss the issue, but I didn't pursue the matter.
Ruud Klaver
AG Projects
Hi Ruud,
I mean, the application uses 10% of cpu without any media active.
Also, pjsua without any params uses less than 1% cpu after start-up,
but 10% after the first call. Is it something you could improve in the
future? Or is it the design, or is it not up to pjproject but to other
code pjrpoject uses? Is it on a Mac only?
-- Best regards,
Alexei
I've noticed exactly the same behaviour on Mac, it's a very annoying
problem. Just having the sound device open seems to consume an inordinate
amount of CPU cycles already. From what I've seen, it actually seems to be a
PortAudio issue. I already went on their mailinglist to discuss the issue,
but I didn't pursue the matter.
Ruud Klaver
AG Projects
I've solved this issure for me, but not shure if this is the most
correct way. It's only about adding one line into pjsua_media.c and
setting snd_auto_close_time=0. I've written about it here:
http://lists.pjsip.org/pipermail/pjsip_lists.pjsip.org/2008-November/005341.html
Alexei Kuznetsov
Hi Alexei,
On 10 Nov 2008, at 12:26, Alexei Kuznetsov wrote:
Hi Ruud,
I mean, the application uses 10% of cpu without any media active.
Also, pjsua without any params uses less than 1% cpu after start-up,
but 10% after the first call. Is it something you could improve in
the
future? Or is it the design, or is it not up to pjproject but to
other
code pjrpoject uses? Is it on a Mac only?
-- Best regards,
Alexei
I've noticed exactly the same behaviour on Mac, it's a very annoying
problem. Just having the sound device open seems to consume an
inordinate
amount of CPU cycles already. From what I've seen, it actually
seems to be a
PortAudio issue. I already went on their mailinglist to discuss the
issue,
but I didn't pursue the matter.
Ruud Klaver
AG Projects
I've solved this issure for me, but not shure if this is the most
correct way. It's only about adding one line into pjsua_media.c and
setting snd_auto_close_time=0. I've written about it here:
http://lists.pjsip.org/pipermail/pjsip_lists.pjsip.org/2008-November/005341.html
Alexei Kuznetsov
I'm not using PJSUA, so this is not a solution for me.
I don't think it is a solution even, this just seems to close the
sound device again if it's idle. The problem still persists, because
it will still consume those CPU cycles while in a call (on top of the
echo cancellation that is). I did some tests with PortAudio directly,
see http://techweb.rfa.org/pipermail/portaudio/2008-October/
008943.html for the relevant thread. The issue wasn't really solved.
Ruud Klaver
AG Projects