OK, yet another thread is closing, where I am the only poster :). For the record, I find out that starting asterisk with -p option (realtime) gives the following table, which now makes more sense. But what is surprising for me is that the load of the server was close to 0 in my original post, thus I would expect similar results. So I'll assume that the previous translation table (without -p) does not show the real capabilities of this server. Lesson learned: start asterisk with -p for close-to-actual translation values.

        g723   gsm  ulaw  alaw  g726 adpcm  slin lpc10  g729 speex  ilbc
  g723     -     -     -     -     -     -     -     -     -     -     -
   gsm     -     -     2     2     2     2     1     3     -     -    13
  ulaw     -     3     -     1     2     2     1     3     -     -    13
  alaw     -     3     1     -     2     2     1     3     -     -    13
  g726     -     3     2     2     -     2     1     3     -     -    13
 adpcm     -     3     2     2     2     -     1     3     -     -    13
  slin     -     2     1     1     1     1     -     2     -     -    12
 lpc10     -     4     3     3     3     3     2     -     -     -    14
  g729     -     -     -     -     -     -     -     -     -     -     -
 speex     -     -     -     -     -     -     -     -     -     -     -
  ilbc     -     4     3     3     3     3     2     4     -     -     -

Cheers,
Soner

----- Original Message ----- From: "Soner Tari" <[EMAIL PROTECTED]> To: "Asterisk Users Mailing List - Non-Commercial Discussion" <[email protected]>
Sent: Thursday, October 20, 2005 9:34 PM
Subject: [Asterisk-Users] slow translations for ilbc and lpc10 on x86_64


Hi All,

When I do 'show translation' on a Linux asterisk 2.6.9-11.EL #1 Wed Jun 8 16:40:06 CDT 2005 x86_64 x86_64 x86_64 GNU/Linux
and
Asterisk CVS HEAD built by [EMAIL PROTECTED] on a x86_64 running Linux on 2005-10-19 19:07:08 UTC

I have very strange lpc10 and ilbc rows (sorry the columns are mixed up, I don't want to use html):

        g723   gsm  ulaw  alaw  g726 adpcm  slin lpc10  g729 speex  ilbc
  g723     -     -     -     -     -     -     -     -     -     -     -
   gsm     -     -     2     2     2     2     1     3     -     -    14
  ulaw     -     3     -     1     2     2     1     3     -     -    14
  alaw     -     3     1     -     2     2     1     3     -     -    14
  g726     -     3     2     2     -     2     1     3     -     -    14
 adpcm     -     3     2     2     2     -     1     3     -     -    14
  slin     -     2     1     1     1     1     -     2     -     -    13
 lpc10     -    48    47    47    47    47    46     -     -     -    59
  g729     -     -     -     -     -     -     -     -     -     -     -
 speex     -     -     -     -     -     -     -     -     -     -     -
  ilbc     -    35    34    34    34    34    33    35     -     -     -

But, even on a Linux asterisk 2.6.9-11.EL #1 Wed Jun 8 16:59:52 CDT 2005 i686 athlon i386 GNU/Linux
and
Asterisk CVS HEAD built by [EMAIL PROTECTED] on a i686 running Linux on 2005-09-16 19:34:57 UTC
(which is a much slower machine, Athlon 2000+, 32bit anyway)
I have much better result on rows for iLBC and lpc10:

        g723   gsm  ulaw  alaw  g726 adpcm  slin lpc10  g729 speex  ilbc
  g723     -     -     -     -     -     -     -     -     -     -     -
   gsm    -     -     2     2     4     2     1     5    -     -    25
  ulaw    -     5     -     1     4     2     1     5    -     -    25
  alaw    -     5     1     -     4     2     1     5    -     -    25
  g726    -     7     4     4     -     4     3     7    -     -    27
 adpcm    -     5     2     2     4     -     1     5    -     -    25
  slin    -     4     1     1     3     1     -     4    -     -    24
 lpc10    -     8     5     5     7     5     4     -    -     -    28
  g729     -     -     -     -     -     -     -     -     -     -     -
 speex     -     -     -     -     -     -     -     -     -     -     -
  ilbc    -     8     5     5     7     5     4     8    -     -     -

I've checked asterisk/codecs/ilbc and lpc10 dirs to see any optimizations for x86_64, but could not find anything.

Should I compile with some different flags? What could be wrong? Or x86_64 is not a supported platform? Any pointers please?

Thanks in advance,
Soner

_______________________________________________
--Bandwidth and Colocation sponsored by Easynews.com --

Asterisk-Users mailing list
[email protected]
http://lists.digium.com/mailman/listinfo/asterisk-users
To UNSUBSCRIBE or update options visit:
http://lists.digium.com/mailman/listinfo/asterisk-users

_______________________________________________
--Bandwidth and Colocation sponsored by Easynews.com --

Asterisk-Users mailing list
[email protected]
http://lists.digium.com/mailman/listinfo/asterisk-users
To UNSUBSCRIBE or update options visit:
  http://lists.digium.com/mailman/listinfo/asterisk-users

Reply via email to