Hi all,
I am writting some new code to the USRP board using gnuradio and
sometimes I
am getting the “uUuU” message, which I understand means that I am not
feeding the usrp.sink module fast enough (right?)
Is there a way to check what is the current CPU load?
Thank you!
On Thu, Apr 17, 2008 at 10:58:31AM -0400, Wireless Monster wrote:
Hi all,
I am writting some new code to the USRP board using gnuradio and sometimes I
am getting the “uUuU” message, which I understand means that I am not
feeding the usrp.sink module fast enough (right?)
Is there a way to check what is the current CPU load?
On a linux system? Try ‘top’, or a graphical tool that does the same
(KDE System Guard, for example).
mb
–
Martin B.
Institut fuer Nachrichtentechnik
Universitaet Karlsruhe
http://www.int.uni-karlsruhe.de
Thank Martin,
However I was thinking in a way to measure the load of each gnuradio
block.
Any clue?
Rgds,
On Thu, Apr 17, 2008 at 12:00 PM, Martin B.
[email protected]
On Fri, Apr 18, 2008 at 10:21 AM, Wireless Monster
[email protected] wrote:
Thank Martin,
However I was thinking in a way to measure the load of each gnuradio block.
Any clue?
Rgds,
I know Eric likes to use oprofile:
http://oprofile.sourceforge.net/about/
You will definitely get a real sense as to what’s eating up your
cycles, but it may be a bit too granular for most people. I have
never used it, personally.
Brian
On Fri, Apr 18, 2008 at 10:25:07AM -0400, Brian P. wrote:
You will definitely get a real sense as to what’s eating up your
cycles, but it may be a bit too granular for most people. I have
never used it, personally.
Brian
It breaks it down per function, plenty fine for most uses.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On Apr 19, 2008, at 2:24 PM, Eric B. wrote:
About OProfile
You will definitely get a real sense as to what’s eating up your
cycles, but it may be a bit too granular for most people. I have
never used it, personally.
Brian
It breaks it down per function, plenty fine for most uses.
You can also do per line.
UNfortunately, the call-graph method is broken on 64-bit linux (AFAI
can tell from really poor documentation and mailing lists), so if two
blocks use the same subroutines (atan2f, fft methods, etc) you can’t
distinguish CPU usage by block.
- -Dan
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.8 (Darwin)
iEYEARECAAYFAkgKaJYACgkQy9GYuuMoUJ6LsQCgkN51+lBziZAgwe9yPpTj7TgS
WXgAnAtjZzjIjC2k0G9B/B2Zdgro2SXT
=n0SP
-----END PGP SIGNATURE-----