Libboost_thread-gcc43-mt-1_36.so.1.36.0

GNU Radio
1

I have followed instructions to download boost
1.36

gnuradio compliled smoothly
I’ve got this problem now: the gnuradio applications I had previously
written compile flawlessly but, when the python script implementing the
application is started
I get:

from gnuradio import gr, eng_notation
File
“/usr/local/lib64/python2.5/site-packages/gnuradio/gr/init.py”,
line 43, in
from gnuradio_swig_python import *
File
“/usr/local/lib64/python2.5/site-packages/gnuradio/gr/gnuradio_swig_python.py”,
line 23, in
from gnuradio_swig_py_runtime import *
File
“/usr/local/lib64/python2.5/site-packages/gnuradio/gr/gnuradio_swig_py_runtime.py”,
line 6, in
import _gnuradio_swig_py_runtime
ImportError: libboost_thread-gcc43-mt-1_36.so.1.36.0: cannot open shared
object file: No such file or directory
FAIL: run_tests

which currently I cannot understand…

what have I done wrong?

thanks

vincenzo

PS.
Eric, also thanks for the precious info about the PS3 chipset.
I think I’ll wait for the USRP2 with its Gigabit Ethernet. Is Gb Eth.
fine
on the PS3?

2

Ok, sorry for bothering.

I understood my mistake by myself. I had forgotten to set the
LD_LIBRARY_PATH environmental variable to the $BOOST__PREFIX/lib.

Now my question is:
I noticed that the load generated by my application is balanced between
the
two cores of my machine… is there any way to go one step back (like a
compiling option or something) and tell the system not to do so but
concentrate the whole load just on one core?

alternatively how can I get the previous trunk version, the one before
introducing dependency on boost 1.35?

thanks

vincenzo

2008/9/3 Vincenzo P. [email protected]

3

On Wed, Sep 03, 2008 at 01:47:00AM +0200, Vincenzo P. wrote:

alternatively how can I get the previous trunk version, the one before
introducing dependency on boost 1.35?

You can get the old behavior by setting the GR_SCHEDULER environment
variable:

$ export GR_SCHEDULER=STS # single-threaded-scheduler (old)
$ export GR_SCHEDULER=TPB # thread-per-block (new)

Eric

4

On Wed, Sep 03, 2008 at 11:55:02AM +0200, Vincenzo P. wrote:

PS
I’ll provide detailed feedback on the oddities I’m currently having as soon
as I’ll have been able to understand what is going on… :slight_smile:

thanks again

The merge that contained the mp-sched was changeset [9336].
To get the pre-merge version:

$ svn co http://gnuradio.org/svn/gnuradio/trunk@9335

Eric