Re: [Discuss-gnuradio] PS3 - gcell_fft
On Fri, Jan 29, 2010 at 03:31:27PM +0100, Thilo Mönicke wrote:
> Hi,
>
> I tryed to set the fft lenght to 16384... but the gcell fft only
> accepts values to 4096. Its defined in the C-Files, but why?
>
> Do you have any working benchmarks for the ps3?
>
> regards,
> Thilo
Sorry Thilo, I steered you in the wrong direction.
The best gcell related FFT code is on CGRAN:
Project description:
https://www.cgran.org/wiki/GcellizedFFTW
Best code:
svn co
https://www.cgran.org/cgran/projects/fftw-gcell/branches/developers/eb/fftw-wip
This code will compute FFT's up to at least 16K points, perhaps 64K (I
don't remember). It's implemented as a modified version of FFTW that
knows about gcell.
To test it, just build and install. By default, let fftw create the
gcell job manager (FFTW will link in the right FFTW spe code. See
fftw-wip/cell/gcell.cc for details.) It's also possible to link this
code into your own SPE executable.
With the modified FFTW library, use the regular gr.fft_vcc block, not
the gcell.fft_vcc block.
Eric
> 2010/1/27 Eric Blossom :
> > On Wed, Jan 27, 2010 at 11:36:14AM +0100, Thilo Mönicke wrote:
> >> Hi,
> >>
> >> I tried to set up a little Testprogram for the PS3 which should show
> >> the advantage of the Power PC structure. For that, the program should
> >> calculate 1 FFTs and take the time. Unfortunately this program is
> >> much slower as a same program running at a GPP. With the tool spu-top
> >> I verified that all SPUs are working...
> >>
> >> Do you have any idea how I could speed up my program?
> >
> > Currently, the break even point is higher than 1024 points due to
> > the overhead to get to the SPEs and back.
> >
> > Try 16384 points :-)
> >
> > Eric
> >
> >
> >> Here is my Python-code (only the intresting part...):
> >>
> >> class BenchmarkFFT(gr.top_block):
> >> def __init__(self):
> >>
> >> ph =
> >> gcell.program_handle_from_filename("/home/moenicke/gnuradio-3.2.2/gcell/lib/spu/gcell_all")
> >> opts = gcell.jm_options(ph,0)
> >> self.mgr = gcell.job_manager(opts)
> >> gcell.set_singleton(self.mgr)
> >>
> >> gr.top_block.__init__(self)
> >>
> >> fft_len=1024
> >> n_ffts=1
> >>
> >> src = gr.noise_source_c(gr.GR_GAUSSIAN, 1.0)
> >> head = gr.head(gr.sizeof_gr_complex, fft_len * n_ffts)
> >> s2v = gr.stream_to_vector(gr.sizeof_gr_complex, fft_len)
> >> fft= gcell.fft_vcc(fft_len, True, [])
> >> sink = gr.null_sink(gr.sizeof_gr_complex * fft_len)
> >>
> >> self.connect (src, head, s2v, fft, sink)
> >>
> >> reguards
> >>
> >> Thilo
> >>
> >> @Eric: thank you for your last answer, it was a big help for me ...
> >
>
___
Discuss-gnuradio mailing list
Discuss-gnuradio@gnu.org
http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 - gcell_fft
Hi,
I tryed to set the fft lenght to 16384... but the gcell fft only
accepts values to 4096. Its defined in the C-Files, but why?
Do you have any working benchmarks for the ps3?
regards,
Thilo
2010/1/27 Eric Blossom :
> On Wed, Jan 27, 2010 at 11:36:14AM +0100, Thilo Mönicke wrote:
>> Hi,
>>
>> I tried to set up a little Testprogram for the PS3 which should show
>> the advantage of the Power PC structure. For that, the program should
>> calculate 1 FFTs and take the time. Unfortunately this program is
>> much slower as a same program running at a GPP. With the tool spu-top
>> I verified that all SPUs are working...
>>
>> Do you have any idea how I could speed up my program?
>
> Currently, the break even point is higher than 1024 points due to
> the overhead to get to the SPEs and back.
>
> Try 16384 points :-)
>
> Eric
>
>
>> Here is my Python-code (only the intresting part...):
>>
>> class BenchmarkFFT(gr.top_block):
>> def __init__(self):
>>
>> ph =
>> gcell.program_handle_from_filename("/home/moenicke/gnuradio-3.2.2/gcell/lib/spu/gcell_all")
>> opts = gcell.jm_options(ph,0)
>> self.mgr = gcell.job_manager(opts)
>> gcell.set_singleton(self.mgr)
>>
>> gr.top_block.__init__(self)
>>
>> fft_len=1024
>> n_ffts=1
>>
>> src = gr.noise_source_c(gr.GR_GAUSSIAN, 1.0)
>> head = gr.head(gr.sizeof_gr_complex, fft_len * n_ffts)
>> s2v = gr.stream_to_vector(gr.sizeof_gr_complex, fft_len)
>> fft= gcell.fft_vcc(fft_len, True, [])
>> sink = gr.null_sink(gr.sizeof_gr_complex * fft_len)
>>
>> self.connect (src, head, s2v, fft, sink)
>>
>> reguards
>>
>> Thilo
>>
>> @Eric: thank you for your last answer, it was a big help for me ...
>
___
Discuss-gnuradio mailing list
Discuss-gnuradio@gnu.org
http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 - gcell_fft
On Wed, Jan 27, 2010 at 11:36:14AM +0100, Thilo Mönicke wrote:
> Hi,
>
> I tried to set up a little Testprogram for the PS3 which should show
> the advantage of the Power PC structure. For that, the program should
> calculate 1 FFTs and take the time. Unfortunately this program is
> much slower as a same program running at a GPP. With the tool spu-top
> I verified that all SPUs are working...
>
> Do you have any idea how I could speed up my program?
Currently, the break even point is higher than 1024 points due to
the overhead to get to the SPEs and back.
Try 16384 points :-)
Eric
> Here is my Python-code (only the intresting part...):
>
> class BenchmarkFFT(gr.top_block):
> def __init__(self):
>
> ph =
> gcell.program_handle_from_filename("/home/moenicke/gnuradio-3.2.2/gcell/lib/spu/gcell_all")
> opts = gcell.jm_options(ph,0)
> self.mgr = gcell.job_manager(opts)
> gcell.set_singleton(self.mgr)
>
> gr.top_block.__init__(self)
>
> fft_len=1024
> n_ffts=1
>
> src = gr.noise_source_c(gr.GR_GAUSSIAN, 1.0)
> head = gr.head(gr.sizeof_gr_complex, fft_len * n_ffts)
> s2v = gr.stream_to_vector(gr.sizeof_gr_complex, fft_len)
> fft= gcell.fft_vcc(fft_len, True, [])
> sink = gr.null_sink(gr.sizeof_gr_complex * fft_len)
>
> self.connect (src, head, s2v, fft, sink)
>
> reguards
>
> Thilo
>
> @Eric: thank you for your last answer, it was a big help for me ...
___
Discuss-gnuradio mailing list
Discuss-gnuradio@gnu.org
http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
[Discuss-gnuradio] PS3 - gcell_fft
Hi,
I tried to set up a little Testprogram for the PS3 which should show
the advantage of the Power PC structure. For that, the program should
calculate 1 FFTs and take the time. Unfortunately this program is
much slower as a same program running at a GPP. With the tool spu-top
I verified that all SPUs are working...
Do you have any idea how I could speed up my program?
Here is my Python-code (only the intresting part...):
class BenchmarkFFT(gr.top_block):
def __init__(self):
ph =
gcell.program_handle_from_filename("/home/moenicke/gnuradio-3.2.2/gcell/lib/spu/gcell_all")
opts = gcell.jm_options(ph,0)
self.mgr = gcell.job_manager(opts)
gcell.set_singleton(self.mgr)
gr.top_block.__init__(self)
fft_len=1024
n_ffts=1
src = gr.noise_source_c(gr.GR_GAUSSIAN, 1.0)
head = gr.head(gr.sizeof_gr_complex, fft_len * n_ffts)
s2v = gr.stream_to_vector(gr.sizeof_gr_complex, fft_len)
fft= gcell.fft_vcc(fft_len, True, [])
sink = gr.null_sink(gr.sizeof_gr_complex * fft_len)
self.connect (src, head, s2v, fft, sink)
reguards
Thilo
@Eric: thank you for your last answer, it was a big help for me ...
___
Discuss-gnuradio mailing list
Discuss-gnuradio@gnu.org
http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 USB chipset
On Fri, Aug 29, 2008 at 05:31:21PM +0200, Vincenzo Pellegrini wrote: > Hi everybody, > I'm planning to begin to learn and explore the fantastic territories of > parallel computing applied to software radio that Eric is currently opening > to us through his work on the Cell BE platform. > (just by the way, thanks Eric!) You're welcome. > Therefore, I will save some money to buy a PS 3 :-) > I know It's not gonna be a easy way to go, but such a computational power is > simply an incredible dream.. :-) > > So, my question: has anybody yet tested the outwards (host to USRP) > throughput of the PS3 usb chipset? > Is it able to cope with our fatal 32MB/s ? It'll do 32MB/s inbound, however the fastest I've been able to get outbound is 25.6MB/s (-i 20). I just tested on a PS3 running Fedora 8 with a 2.6.25.9-40.fc8 kernel. The PS3 firmware was updated about a month ago. > Thanks > Vincenzo > > PS. > Is anybody using an adapter to connect the PS3 video output to a standard > VGA monitor? I tried pretty much all the tricks, and after finding none of them satisfactory, bought a low cost HDMI compatible monitor. FWIW, I bought a BENQ FP222W. I tried: (1) HDMI to DVI cable. Doesn't work, PS3 requires HDCP DRM cruft (2) Composite video into flat panel (the yellow RCA connector): barely readable enough to get it to boot. (3) Gadget that converted component output to VGA: barely readable, didn't support the 480i mode that it booted up in, thus had to change between kludge (2) and kludge (3) on the fly by plugging and unplugging cables. (4) Finally spend about $250 to buy an HDMI monitor. Eric ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
[Discuss-gnuradio] PS3 USB chipset
Hi everybody, I'm planning to begin to learn and explore the fantastic territories of parallel computing applied to software radio that Eric is currently opening to us through his work on the Cell BE platform. (just by the way, thanks Eric!) Therefore, I will save some money to buy a PS 3 :-) I know It's not gonna be a easy way to go, but such a computational power is simply an incredible dream.. :-) So, my question: has anybody yet tested the outwards (host to USRP) throughput of the PS3 usb chipset? Is it able to cope with our fatal 32MB/s ? Thanks Vincenzo PS. Is anybody using an adapter to connect the PS3 video output to a standard VGA monitor? -- Vincenzo Pellegrini ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 cross-compile problem - SOLVED
On Sun, Jul 27, 2008 at 03:32:22PM +0200, Andreas Fernström wrote: > > Problem solved! > > I just had to add /opt/cell/toolchain/bin > to PATH on the build machine. > > After that, configure-cell-cross, make, 'make check' and 'make install' > worked without problems. > > I would however like to point out some things that maybe should be in the > wiki: login:guest password: gnuradio > Next thing would be to get the usrp connected and running, but first > I'd like to have some sound and display. Anyone that have "been > there, done that" that could point me in the right direction to get > those two things working on the PS3, or presented remotely?? Sounds works out of the box if you using -I or -O plughw:0,0 on the audio examples. If you want to help make hw:0,0 work drop me a line. I'll tell you what needs to happen. For a display, just ssh in to the ps3 from your build host using the -X option to enable X11 forwarding. Eric ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 cross-compile problem - SOLVED
Problem solved!
I just had to add /opt/cell/toolchain/bin
to PATH on the build machine.
After that, configure-cell-cross, make, 'make check' and 'make install' worked
without problems.
I would however like to point out some things that maybe should be in the wiki:
1. There is a line that says: Download the latest Cell
http://www.kernel.org/pub/linux/kernel/people/geoff/cell/CELL-Linux-CL_20080201-ADDON.iso
(ADDON CD iso).
The link doesn't work, I used an ADDON cd at
http://www.kernel.org/pub/linux/kernel/people/geoff/cell/ps3-distro-kit/
2. There is a line that says: At the kboot command prompt, enter: linux64
xdriver=fbdev video=720p install text.
I'm not using the PS3 for games, so I have connected a 20+ years old TV to it,
which (probably combined with being European) does not handle the 720p
resolution. I had to use video=576i instead. For guidance look at
http://manuals.playstation.net/document/en/ps3/current/settings/videooutput.html.
3. You HAVE to use the /mnt/cell-root location on your build machine when
mounting the ps3 root. You really do.
I tried some other location first, but even if you change the variables in the
scripts, I found that /mnt/cell-root was hard linked in some places. When
trying to change mountpoint to /mnt/cell-root, which I should have used the
first time, I noticed that running tweak-cell-for-cross-compiling a second time
with another location isn't wise. The result is additive, that is, if you first
got the .so files tweaked to path1/lib...so, the second time you get
path2/path1/lib...so. Not good. So do use /mnt/cell-root as it says.
4. Read the error messages. Read the error messages. Read the error message...
Get it?
If it says 'ppu32-gcc: command not found', that means that the ppu32-gcc
command was not found. Do you have the file? Where is it located? Is the PATH
ok? Are you checking the PATH on the right machine?
I had missed that ppu32-gcc had to be accessible with a good PATH on the build
machine too.
Next thing would be to get the usrp connected and running, but first I'd like
to have some sound and display. Anyone that have "been there, done that" that
could point me in the right direction to get those two things working on the
PS3, or presented remotely??
Andreas
> From: [EMAIL PROTECTED]
> To: discuss-gnuradio@gnu.org
> Date: Sun, 27 Jul 2008 04:02:32 +0200
> Subject: [Discuss-gnuradio] PS3 cross-compile problem
>
>
> Hi!
>
> I have problem cross compiling GNU Radio on my x86_64 machine for my ps3.
> I've never tried cross compiling before, so I'm not sure that I've got
> everything right,
> but when I try (on the x86_64 machine) to run
>
> [EMAIL PROTECTED] trunk]$ ./configure-cell-cross
> --prefix=/mnt/share/cell-install
>
> I get the following errors printed in config.log
>
> configure:3036: checking for powerpc64-unknown-linux-gnu-gcc
> configure:3063: result: ppu32-gcc
> configure:3341: checking for C compiler version
> configure:3348: ppu32-gcc --version>&5
> ./configure: line 3349: ppu32-gcc: command not found
> configure:3351: $? = 127
> configure:3358: ppu32-gcc -v>&5
> ./configure: line 3359: ppu32-gcc: command not found
> configure:3361: $? = 127
> configure:3368: ppu32-gcc -V>&5
> ./configure: line 3369: ppu32-gcc: command not found
> configure:3371: $? = 127
> configure:3394: checking for C compiler default output file name
> configure:3421: ppu32-gcc -I/mnt/cell-root/usr/include -L/mnt/cell-root/lib
> -L/mnt/cell-root/usr/lib conftest.c>&5
> ./configure: line 3422: ppu32-gcc: command not found
> configure:3424: $? = 127
> configure:3462: result:
> configure: failed program was:
> | /* confdefs.h. */
> | #define PACKAGE_NAME ""
> | #define PACKAGE_TARNAME ""
> | #define PACKAGE_VERSION ""
> | #define PACKAGE_STRING ""
> | #define PACKAGE_BUGREPORT ""
> | #define PACKAGE "gnuradio"
> | #define VERSION "3.1svn"
> | /* end confdefs.h. */
> |
> | int
> | main ()
> | {
> |
> | ;
> | return 0;
> | }
> configure:3469: error: C compiler cannot create executables
>
>
>
> It seems that there is some problem with the ppu32-gcc setting...
>
> so I checked (the ps3) but got:
>
> [EMAIL PROTECTED] ~]$ ppu32-gcc -v
> Using built-in specs.
> Target: ppu
> Configured with: ../toolchain/gcc/configure --prefix=/usr
> --mandir=/usr/share/man --infodir=/usr/share/info --with-as=/usr/bin/ppu-as
> --with-ld=/usr/bin/ppu-ld --enable-threads --with-system-zlib
> --disable-checking --enable-__cxa_atexit --disable-libunwind-exceptions
> --enable-languages=c,c++,fortran,ada --disable-nls
[Discuss-gnuradio] PS3 cross-compile problem
Hi!
I have problem cross compiling GNU Radio on my x86_64 machine for my ps3.
I've never tried cross compiling before, so I'm not sure that I've got
everything right,
but when I try (on the x86_64 machine) to run
[EMAIL PROTECTED] trunk]$ ./configure-cell-cross
--prefix=/mnt/share/cell-install
I get the following errors printed in config.log
configure:3036: checking for powerpc64-unknown-linux-gnu-gcc
configure:3063: result: ppu32-gcc
configure:3341: checking for C compiler version
configure:3348: ppu32-gcc --version>&5
./configure: line 3349: ppu32-gcc: command not found
configure:3351: $? = 127
configure:3358: ppu32-gcc -v>&5
./configure: line 3359: ppu32-gcc: command not found
configure:3361: $? = 127
configure:3368: ppu32-gcc -V>&5
./configure: line 3369: ppu32-gcc: command not found
configure:3371: $? = 127
configure:3394: checking for C compiler default output file name
configure:3421: ppu32-gcc -I/mnt/cell-root/usr/include -L/mnt/cell-root/lib
-L/mnt/cell-root/usr/lib conftest.c >&5
./configure: line 3422: ppu32-gcc: command not found
configure:3424: $? = 127
configure:3462: result:
configure: failed program was:
| /* confdefs.h. */
| #define PACKAGE_NAME ""
| #define PACKAGE_TARNAME ""
| #define PACKAGE_VERSION ""
| #define PACKAGE_STRING ""
| #define PACKAGE_BUGREPORT ""
| #define PACKAGE "gnuradio"
| #define VERSION "3.1svn"
| /* end confdefs.h. */
|
| int
| main ()
| {
|
| ;
| return 0;
| }
configure:3469: error: C compiler cannot create executables
It seems that there is some problem with the ppu32-gcc setting...
so I checked (the ps3) but got:
[EMAIL PROTECTED] ~]$ ppu32-gcc -v
Using built-in specs.
Target: ppu
Configured with: ../toolchain/gcc/configure --prefix=/usr
--mandir=/usr/share/man --infodir=/usr/share/info --with-as=/usr/bin/ppu-as
--with-ld=/usr/bin/ppu-ld --enable-threads --with-system-zlib
--disable-checking --enable-__cxa_atexit --disable-libunwind-exceptions
--enable-languages=c,c++,fortran,ada --disable-nls --enable-clocale=gnu
--enable-version-specific-runtime-libs --with-long-double-128
--program-prefix=ppu- --disable-bootstrap --host=ppu
--build=powerpc64-unknown-linux-gnu --target=ppu
Thread model: posix
gcc version 4.1.1
Can anyone tell me what is wrong??
Andreas
___
Discuss-gnuradio mailing list
Discuss-gnuradio@gnu.org
http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 installation
On Thu, Jul 24, 2008 at 06:41:29AM -0700, Walter Snel wrote: > > Radio afficionados, > > I'm trying to compile the normal gnu-radio trunk on the PS3. Unfortunately I > run into an issue. > > I'd appreciate a pointer in the right direction. > > I used FC9, installed the cell SDK (3.0.0.1.0) and compiled the examples, so > far, so good. > > However, when compiling gnuradio I run into a compilation error in the gcell > section. It indicates "ppu_intrinsics.h" is not found. On developerworks a > similar question led to a reply indicating ppu-gcc instead of gcc should be > used. Is this a known build issue ? > > Any help is appreciated, > Walter To build on the PS3 and use gcell, you're going to need what we're calling the "cross compilation environment", even though you're not cross compiling :-) Yes, you'll need to use ppu-g++ (or grab a copy of ppu_intrinsics.h from somewhere and install it where we'll find it). Please take a look at the configure-cell-cross shell script in the top directory. To avoid waiting forever for your compilations, I suggest you cross compile from a typical x86 box. See http://gnuradio.org/trac/wiki/CrossCompilingForCell Also, at least on x86 and x86_64 we've found that Fedora 9 is undesirable because of the buggy version of g++ that they ship. Our QA code fails, and so does that of other projects. Eric > - compile error > - > /bin/sh ../../../../libtool --tag=CXX --mode=compile g++ -DHAVE_CONFIG_H > -I. -I../../../.. -DOMNITHREAD_POSIX=1 -I/home/wsnel/trunk/omnithread > -I/home/wsnel/trunk/mblock/src/lib -I/home/wsnel/trunk/gcell/src/include > > -I/home/wsnel/trunk/gcell/src/lib/runtime > -I/home/wsnel/trunk/gcell/src/lib/general > -I/home/wsnel/trunk/gcell/src/lib/wrapper > -I../../../../gcell/src/ibm/sync/ppu_source -g -O2 -Wall > -Woverloaded-virtual -pthread -MT gc_job_manager_impl.lo -MD -MP -MF > .deps/gc_job_manager_impl.Tpo -c -o gc_job_manager_impl.lo > gc_job_manager_impl.cc > g++ -DHAVE_CONFIG_H -I. -I../../../.. -DOMNITHREAD_POSIX=1 > -I/home/wsnel/trunk/omnithread -I/home/wsnel/trunk/mblock/src/lib > -I/home/wsnel/trunk/gcell/src/include > -I/home/wsnel/trunk/gcell/src/lib/runtime > -I/home/wsnel/trunk/gcell/src/lib/general > -I/home/wsnel/trunk/gcell/src/lib/wrapper > -I../../../../gcell/src/ibm/sync/ppu_source -g -O2 -Wall > -Woverloaded-virtual -pthread -MT gc_job_manager_impl.lo -MD -MP -MF > .deps/gc_job_manager_impl.Tpo -c gc_job_manager_impl.cc -fPIC -DPIC -o > .libs/gc_job_manager_impl.o > In file included from gc_job_manager_impl.cc:32: > ../../../../gcell/src/ibm/sync/ppu_source/atomic_dec_if_positive.h:44:28: > error: ppu_intrinsics.h: No such file or directory > In file included from > ../../../../gcell/src/ibm/sync/ppu_source/atomic_dec_if_positive.h:46, > from gc_job_manager_impl.cc:32: > - compile error > - > -- ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
[Discuss-gnuradio] PS3 installation
Radio afficionados, I'm trying to compile the normal gnu-radio trunk on the PS3. Unfortunately I run into an issue. I'd appreciate a pointer in the right direction. I used FC9, installed the cell SDK (3.0.0.1.0) and compiled the examples, so far, so good. However, when compiling gnuradio I run into a compilation error in the gcell section. It indicates "ppu_intrinsics.h" is not found. On developerworks a similar question led to a reply indicating ppu-gcc instead of gcc should be used. Is this a known build issue ? Any help is appreciated, Walter - compile error - /bin/sh ../../../../libtool --tag=CXX --mode=compile g++ -DHAVE_CONFIG_H -I. -I../../../.. -DOMNITHREAD_POSIX=1 -I/home/wsnel/trunk/omnithread -I/home/wsnel/trunk/mblock/src/lib -I/home/wsnel/trunk/gcell/src/include -I/home/wsnel/trunk/gcell/src/lib/runtime -I/home/wsnel/trunk/gcell/src/lib/general -I/home/wsnel/trunk/gcell/src/lib/wrapper -I../../../../gcell/src/ibm/sync/ppu_source -g -O2 -Wall -Woverloaded-virtual -pthread -MT gc_job_manager_impl.lo -MD -MP -MF .deps/gc_job_manager_impl.Tpo -c -o gc_job_manager_impl.lo gc_job_manager_impl.cc g++ -DHAVE_CONFIG_H -I. -I../../../.. -DOMNITHREAD_POSIX=1 -I/home/wsnel/trunk/omnithread -I/home/wsnel/trunk/mblock/src/lib -I/home/wsnel/trunk/gcell/src/include -I/home/wsnel/trunk/gcell/src/lib/runtime -I/home/wsnel/trunk/gcell/src/lib/general -I/home/wsnel/trunk/gcell/src/lib/wrapper -I../../../../gcell/src/ibm/sync/ppu_source -g -O2 -Wall -Woverloaded-virtual -pthread -MT gc_job_manager_impl.lo -MD -MP -MF .deps/gc_job_manager_impl.Tpo -c gc_job_manager_impl.cc -fPIC -DPIC -o .libs/gc_job_manager_impl.o In file included from gc_job_manager_impl.cc:32: ../../../../gcell/src/ibm/sync/ppu_source/atomic_dec_if_positive.h:44:28: error: ppu_intrinsics.h: No such file or directory In file included from ../../../../gcell/src/ibm/sync/ppu_source/atomic_dec_if_positive.h:46, from gc_job_manager_impl.cc:32: - compile error - -- View this message in context: http://www.nabble.com/Re%3A-gcell-status--tp18038915p18578479.html Sent from the GnuRadio mailing list archive at Nabble.com. ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] ps3 cross compile with ps3 nfs server
On Tue, Jun 17, 2008 at 11:18:08AM -0600, Chuck Swiger wrote: > You *can* cross compile using the latest ps3 slim & trim 2.6.25.4 kernel > from CELL-Linux-CL_20080609-ADDON.iso by using samba. Just install it with > "yum install samba" and make the ps3 root export in /etc/samba/smb.conf on > host ps3: > > [cell-root] > path = / > public = yes > writable = no > printable = no > guest ok = yes > > add a user with smbpasswd > > # smbpasswd -a chuck > > > start it > > # /etc/init.d/smb start > > and mount it on the build machine: > > # mount -t cifs //ps3/cell-root /mnt/cell-root -o username=chuck > > > proceed are normal. Thanks for letting us know! Eric ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
[Discuss-gnuradio] ps3 cross compile with ps3 nfs server
You *can* cross compile using the latest ps3 slim & trim 2.6.25.4 kernel from CELL-Linux-CL_20080609-ADDON.iso by using samba. Just install it with "yum install samba" and make the ps3 root export in /etc/samba/smb.conf on host ps3: [cell-root] path = / public = yes writable = no printable = no guest ok = yes add a user with smbpasswd # smbpasswd -a chuck start it # /etc/init.d/smb start and mount it on the build machine: # mount -t cifs //ps3/cell-root /mnt/cell-root -o username=chuck proceed are normal. ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
[Discuss-gnuradio] ps3 kernel upgrade
Eric et al - Obtained and installed CELL-Linux-CL_20080609-ADDON.iso with the kernel 2.6.25.4, bad news is it still does not have the nfs modules. The good news is with 220MB free memory compiling on the ps3 isn't too slow - I had two windows compiling different parts and swap was rarely being used. The bad news is make check still segfaults on qa_gcell_wrapper. The good news is the offload stuff still basically works and playing with benchmark_nop really makes sense now. Running 1000uSec of nops per thread, using 1000 threads on one SPU takes just about a second. Running 2000 threads on 2 SPU takes just about a second. 6000 threads on 6 SPU takes about a second. Using 500uSec of nop, 12000 threads on 6 SPU takes about one seconds, etc, etc, etc. The benchmark_* examples look like a good jumping off place for using the gcell tools with ones own spu code. --Chuck ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] ps3 gnuradio
On Sat, 2008-05-31 at 13:45 -0400, Charles Swiger wrote:
> 'make check' fails with a segfault in 'gcell/src/apps/test_all'.
played around with gdb and finally found this while stepping:
Breakpoint 17, ~worker_ctx (this=0xf7e0071c) at
gc_job_manager_impl.cc:1241
1241 if (spe_ctx){
(gdb)
1242int r = spe_context_destroy(spe_ctx);
(gdb)
1241 if (spe_ctx){
(gdb)
1242int r = spe_context_destroy(spe_ctx);
(gdb)
1243if (r != 0){
(gdb)
1246spe_ctx = 0;
(gdb)
1248 state = WS_FREE;
(gdb)
1249}
(gdb)
1248 state = WS_FREE;
(gdb)
1249}
(gdb)
~gc_job_manager_impl (this=0xf7e006e8)
at /mnt/cell-root/usr/include/boost/detail/shared_count.hpp:159
159 if( pi_ != 0 ) pi_->release();
(gdb)
74 );
(gdb)
151 if( atomic_decrement( &use_count_ ) == 0 )
(gdb)
335 }
(gdb)
~gc_job_manager (this=0xf7e006e8) at gc_job_manager.cc:58
58 gc_job_manager::~gc_job_manager()
(gdb)
61 }
(gdb)
58 gc_job_manager::~gc_job_manager()
(gdb)
61 }
(gdb)
boost::detail::sp_counted_impl_p::dispose (
this=)
at /mnt/cell-root/usr/include/boost/detail/sp_counted_impl.hpp:76
76 boost::checked_delete( px_ );
(gdb)
qa_gcell_general::generic_test_body (this=,
[EMAIL PROTECTED])
at /mnt/cell-root/usr/include/boost/detail/sp_counted_base_gcc_ppc.hpp:74
74 );
(gdb)
165 if( atomic_decrement( &weak_count_ ) == 0 )
(gdb)
167 destroy();
(gdb)
boost::detail::sp_counted_base::destroy (this=0xf7e17bb0)
at /mnt/cell-root/usr/include/boost/detail/sp_counted_base_gcc_ppc.hpp:134
134 delete this;
(gdb)
~sp_counted_impl_p (this=0xf7e17bb0)
at /mnt/cell-root/usr/include/boost/detail/sp_counted_impl.hpp:52
52 {
(gdb)
122 {
(gdb)
52 {
(gdb)
122 {
(gdb)
52 {
(gdb)
boost::detail::sp_counted_base::destroy (this=0xfa3ca00)
at /mnt/cell-root/usr/include/boost/detail/sp_counted_base_gcc_ppc.hpp:135
135 }
(gdb)
qa_gcell_general::generic_test_body (this=,
[EMAIL PROTECTED])
at /mnt/cell-root/usr/include/boost/shared_ptr.hpp:253
253 BOOST_ASSERT(px != 0);
(gdb)
63if (!mgr->wait_job(jd.get())){
(gdb)
Program received signal SIGSEGV, Segmentation fault.
0x in ?? ()
___
Discuss-gnuradio mailing list
Discuss-gnuradio@gnu.org
http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
[Discuss-gnuradio] ps3 gnuradio
Here's what I found on the ps3 system so far: Current system software is 2.35 and had to use a later 'CELL-Linux-CL-20080201-ADDON.iso' otherwise got a blank screen on 'start other os'. 20071023 is the version in the wiki, I'll try to update things at some time. Kernel 2.6.21 works fine but has the shutdown hang issue. Kernel 2.6.24 reboots ok but is missing nfsd.ko, or I just could not otherwise get nfs to work. Turning off all unnecessary services and modules got 160 out of 217MB free ;) Plow through cross compile, build failed with -j6, but worked with just 'make'. Wiki says -j2. 'make check' fails with a segfault in 'gcell/src/apps/test_all'. Commenting out 'runner.addTest(qa_gcell_wrapper::suite());' makes it work again. Played around with some of the other apps and got this nice chart: http://www.swigerco.com/R-8534-20080531-0810.png Went ahead and did 'make install', set PYTHONPATH and imported gr and audio. Tried dialtone but got 'audio_alsa_sink[hw:0,0]: failed to set access mask: Invalid argument'. Used a file sink and got 'gr_vmcircbuf_createfilemapping: createfilemapping is not available' but got a bunch of output that plots correctly on a big-endian machine. Finally found that 'audio.sink(48000,"plughw:0,0")' works. --Chuck ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3
On Thu, Dec 13, 2007 at 06:35:56PM +0200, Aadil Volkwin wrote: > Hi, > > I've noticed the increased interest in running GnuRadio and the USRP with > the PS3 as the computing platform. > > My naive thought is that this is to explore the performance of the cell > processor architecture. > > Would someone be willing to offer me their thoughts on this choice and what > are the pros and cons, from their experience? > I apologise in advance if its a silly question. Pros: 25 GFLOPs peak per SPE (PS3 has 6 usable SPEs) Cons: different/low-level programming environment. See any of the overview docs on the ibm.com site. Eric ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
[Discuss-gnuradio] PS3
Hi, I've noticed the increased interest in running GnuRadio and the USRP with the PS3 as the computing platform. My naive thought is that this is to explore the performance of the cell processor architecture. Would someone be willing to offer me their thoughts on this choice and what are the pros and cons, from their experience? I apologise in advance if its a silly question. Aadil ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
[Discuss-gnuradio] PS3 Cross Compile
So I went out and got a PS3 to play with. I couldn't resist. I messed around with F7 for a while and then said what the hell, I wanna get SUSE 10.3 on there since you can get all the free tools at the BSC site. I almost have the cross compile environment setup but have been running into some linker/toolchain issues and wondering if anyone might be able to help me if they have ran into this before. I googled it and only got a bit of information. Here is the last part of the make output: libtool: link: warning: library `/mnt/cell-root/usr/lib/libfftw3f.la' was moved. ppu32-g++ -g -O2 -Wall -Woverloaded-virtual -pthread -o .libs/benchmark_dotprod benchmark_dotprod.o -L/mnt/cell-root/lib -L/mnt/cell-root/usr/lib ../../../gnuradio-core/src/lib/.libs/libgnuradio- core.so /mnt/share/gnuradio/omnithread/.libs/libgromnithread.so -lrt /mnt/cell-root/usr/lib/libfftw3f.so -Wl,--rpath -Wl,/mnt/share/cell-install/lib -Wl,--rpath -Wl,/mnt/cell-root/usr/lib /opt/cell/toolchain/bin/ppu-ld: skipping incompatible /mnt/cell-root/usr/lib/librt.so when searching for -lrt /opt/cell/toolchain/bin/ppu-ld: skipping incompatible /mnt/cell-root/usr/lib/libm.so when searching for -lm ../../../gnuradio-core/src/lib/.libs/libgnuradio-core.so: undefined reference to `_dl_hwcap' collect2: ld returned 1 exit status make[4]: *** [benchmark_dotprod] Error 1 make[4]: Leaving directory `/mnt/share/gnuradio/gnuradio-core/src/tests' make[3]: *** [all-recursive] Error 1 make[3]: Leaving directory `/mnt/share/gnuradio/gnuradio-core/src' make[2]: *** [all-recursive] Error 1 make[2]: Leaving directory `/mnt/share/gnuradio/gnuradio-core' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/mnt/share/gnuradio' make: *** [all] Error 2 Now, I made some changes to the configure-cell-cross but only to correct some earlier errors with errno.h, which it did. If someone could just point me in the right direction to fix this that would be awesome. Thanks -- Newell Jensen ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] ps3 performance
This and other questions. You are not doing something right. Just using the spe's for fft's we will see a huge speed up in gnuradio. I can run filter, NCO, baud clock and carrier recovery, and equalizer on the cell on 40 Mbaud signals. Trying getting that out of a (Intel) cola nut. We can't support that data rate into the ps3 but I can into the cell servers at work. The PS3 will run like crap if you are running a windowing environment, even one of the lightweight ones because of small memory. That said, we are going to be able to really increase our signal processing throughput by selective use of the SPE's and decent planning our use. Bob Eric Blossom wrote: On Tue, Oct 30, 2007 at 09:44:39AM +0200, Juha Vierinen wrote: BTW, terrasoft sells ps3 pre-installed with linux (and a root password supplied). I have never even connected a monitor to my ps3. But as to performance, has anyone measured decent performance on a ps3? I am starting to think that it is not possible to achieve anything near the theoretical peak performance. I am often getting better performance with a intel core2 than on the 6+1 processors that come with Cell. Intel c compiler often comes up with a faster program just by autovectorizing a standard c program -- and this is compared to something hand vectorized and paralellized for Cell. It could be just that I'm not at all good at optimizing code, but this is also true for some of the examples that come with the IBM SDK (eg. the euler solver). The only thing that I have found to be impressive was the cell optimized fft. (1) Are you ordering and aligning your SPE instructions such that you're getting dual issue? (2) Are you providing branch hints? A mispredicted branch stalls for 11 cycles. Eric ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio -- AMSAT Director and VP Engineering. Member: ARRL, AMSAT-DL, TAPR, Packrats, NJQRP, QRP ARCI, QCWA, FRC. ARRL SDR WG Chair “An optimist may see a light where there is none, but why must the pessimist always run to blow it out?” Descartes ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] ps3 performance
On Tue, Oct 30, 2007 at 09:44:39AM +0200, Juha Vierinen wrote: > BTW, terrasoft sells ps3 pre-installed with linux (and a root password > supplied). I have never even connected a monitor to my ps3. > > But as to performance, has anyone measured decent performance on a > ps3? I am starting to think that it is not possible to achieve > anything near the theoretical peak performance. I am often getting > better performance with a intel core2 than on the 6+1 processors that > come with Cell. Intel c compiler often comes up with a faster program > just by autovectorizing a standard c program -- and this is compared > to something hand vectorized and paralellized for Cell. It could be > just that I'm not at all good at optimizing code, but this is also > true for some of the examples that come with the IBM SDK (eg. the > euler solver). The only thing that I have found to be impressive was > the cell optimized fft. (1) Are you ordering and aligning your SPE instructions such that you're getting dual issue? (2) Are you providing branch hints? A mispredicted branch stalls for 11 cycles. Eric ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
[Discuss-gnuradio] PS3 transfer rates
Hi! I had a PS3 laying around and decided to try to connect one of my usrps to it... they seem to be able to talk to each other, but what transfer rates am I supposed to expect? I get : [EMAIL PROTECTED] apps]$ ./test_usrp_standard_rx rx_overrun rx_overrun xfered 1.34e+08 bytes in 4.32 seconds. 3.104e+07 bytes/sec. cpu time = 0.9172 noverruns = 41 and [EMAIL PROTECTED] apps]$ ./test_usrp_standard_tx xfered 1.34e+08 bytes in 4.69 seconds. 2.863e+07 bytes/sec. cpu time = 1.15 0 underruns The results are very reproducible. Andreas ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 with freedom
> it really does speak to the foresight (and good taste) of Sony that > [very early] they had published a bootloader and clear > instructions on how to put Linux on the thing. Each and everything they > have done in this area has been with the support of IBM and released > GPL. This is indeed priaseworthy -- particularly given Sony's long and ongoing history of making life hard for free software (e.g. with documentation or drivers for their laptop chips) and their unfortunate fondness for DRM and gratuitous incompatibility in all its forms. > We are not using xlc and are not likely to use xlc. We are using the > free parts of the SDK including spu-gcc and ppu-gcc for example. I'm very glad to hear it. > While I understand your concern, support the basis for your stance > on this and many of your other heartfelt issues turned cause célèbre, it > is my OPINION that in this case, your concerns are misplaced. I'm glad that people of the caliber of you and Eric are watching these issues closely. I'm looking forward to the day when there is a documented way to get this free software -- without wading through reading and agreeing to a pile of nonfree licenses. :-) John ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 versus freedom
Brian Padalino wrote: On 10/9/07, Robert McGwier <[EMAIL PROTECTED]> wrote: ... However, nothing about the gpl prohibits individuals from using xlc to generate binaries for their own use so long as they do not distribute the binary. ... This is more for curiosity than anything, but where do the USRP FPGA bit files fit in with this? Brian released GPL -- AMSAT Director and VP Engineering. Member: ARRL, AMSAT-DL, TAPR, Packrats, NJQRP, QRP ARCI, QCWA, FRC. ARRL SDR WG Chair “An optimist may see a light where there is none, but why must the pessimist always run to blow it out?” Descartes ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 versus freedom
On 10/9/07, Robert McGwier <[EMAIL PROTECTED]> wrote: > ... > However, nothing about the gpl prohibits individuals from using xlc to > generate binaries for their own use so long as they do not distribute > the binary. > ... This is more for curiosity than anything, but where do the USRP FPGA bit files fit in with this? Brian ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 versus freedom
John: However, let me add hastily, nothing irks me more than the hypervisor and the hidden IP in the graphics chip. So be it. I do not want to steal anything that is covered by M$ DRM anyway and I do not envision us using the PS3 to do graphical display (since its memory is too small to support a window manager with any serious capability). I suggest that we do exactly what your note suggests and stick to what the thing is capable of delivering to use with free tools. Bob Robert McGwier wrote: John: We are not using xlc and are not likely to use xlc. We are using the free parts of the SDK including spu-gcc and ppu-gcc for example. For this project in fact, we cannot distribute any code built with xlc. However, nothing about the gpl prohibits individuals from using xlc to generate binaries for their own use so long as they do not distribute the binary. The PS3 has been available to the end users for only a few months. Yet, it really does speak to the foresight (and good taste) of Sony that before you or I could go to a store, even if it meant standing in line for hours and hours, they had published a bootloader and clear instructions on how to put Linux on the thing. Each and everything they have done in this area has been with the support of IBM and released GPL. Sony has multiple representatives in daily conversation with the lead developer for the kernel. Eric and I both monitor the group and the results will be the clear instructions (if non trivial) on bring up a useful kernel on the PS3 that allows us to develop and use gnu tools. IBM has contributed, under the GPL, their optimizations to FFTW. The impact of this on GnuRadio using the PS3 is not to be underestimated. While I understand your concern, support your the basis for your stance on this and many of your other heartfelt issues turned cause célèbre, it is my OPINION that in this case, your concerns are misplaced. Bob McGwier (speaking only for myself) -- AMSAT Director and VP Engineering. Member: ARRL, AMSAT-DL, TAPR, Packrats, NJQRP, QRP ARCI, QCWA, FRC. ARRL SDR WG Chair “An optimist may see a light where there is none, but why must the pessimist always run to blow it out?” Descartes ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 versus freedom
John: We are not using xlc and are not likely to use xlc. We are using the free parts of the SDK including spu-gcc and ppu-gcc for example. For this project in fact, we cannot distribute any code built with xlc. However, nothing about the gpl prohibits individuals from using xlc to generate binaries for their own use so long as they do not distribute the binary. The PS3 has been available to the end users for only a few months. Yet, it really does speak to the foresight (and good taste) of Sony that before you or I could go to a store, even if it meant standing in line for hours and hours, they had published a bootloader and clear instructions on how to put Linux on the thing. Each and everything they have done in this area has been with the support of IBM and released GPL. Sony has multiple representatives in daily conversation with the lead developer for the kernel. Eric and I both monitor the group and the results will be the clear instructions (if non trivial) on bring up a useful kernel on the PS3 that allows us to develop and use gnu tools. IBM has contributed, under the GPL, their optimizations to FFTW. The impact of this on GnuRadio using the PS3 is not to be underestimated. While I understand your concern, support your the basis for your stance on this and many of your other heartfelt issues turned cause célèbre, it is my OPINION that in this case, your concerns are misplaced. Bob McGwier (speaking only for myself) John Gilmore wrote: If you want to run on the PS3, you're most likely going to want the IBM SDK 3.0. The SDK really, really wants FC 7 on the PS3. Some people will do anything for crunchons -- or "promised future" crunchons. (A crunchon is a unit of number-crunching.) I was wondering why this SDK wasn't already part of the SuSe or FC7 releases. Ten minutes of research later, the answer is: it's proprietary. Why would anyone on this mailing list want to install proprietary compilers? The IBM license agreement is extortionate. Besides the usual rape and pillage, it says that: http://www14.software.ibm.com/cgi-bin/weblap/lap.pl?la_formnum=&li_formnum=L-MCHN-6MVMPV&title=XL%20C/C%2B%2B%20Alpha%20Edition%20for%20Cell%20Processor * You are not authorized to use the Program for productive purposes * THE PROGRAM MAY CONTAIN A DISABLING DEVICE THAT WILL PREVENT IT FROM BEING USED AFTER THE EVALUATION PERIOD ENDS. YOU MAY NOT TAMPER WITH THIS DISABLING DEVICE OR THE PROGRAM. YOU SHOULD TAKE PRECAUTIONS TO AVOID ANY LOSS OF DATA THAT MIGHT RESULT WHEN THE PROGRAM CAN NO LONGER BE USED. * You assign to IBM all right, title, and interest (including ownership of copyright) in any data, suggestions, or written materials that 1) are related to the Program and 2) You provide to IBM. * Your right to run the program "ends after 90 days." [I hope Eric hasn't provided IBM a copy of the GNU Radio code that compiles under this compiler: IBM will end up owning the copyright on GNU Radio. This may occur even if Eric just posts the code in a public place where IBM can download it.] If the above wasn't enough, it's also against GNU project policy to use the mailing list or project documentation to advertise or advocate for proprietary software. If the GNU compilers for the CELL aren't good enough for us, I suggest that we improve them. If we just can't exploit the CELL processor without proprietary software and patented algorithms, then I suggest we go back to focusing on hosting GNU Radio on hardware that comes with freedom. John ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio -- AMSAT Director and VP Engineering. Member: ARRL, AMSAT-DL, TAPR, Packrats, NJQRP, QRP ARCI, QCWA, FRC. ARRL SDR WG Chair “An optimist may see a light where there is none, but why must the pessimist always run to blow it out?” Descartes ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 versus freedom
On Mon, Oct 08, 2007 at 09:59:03PM -0700, John Gilmore wrote: > > If you want to run on the PS3, you're most likely going to want the > > IBM SDK 3.0. The SDK really, really wants FC 7 on the PS3. > > Some people will do anything for crunchons -- or "promised future" crunchons. > (A crunchon is a unit of number-crunching.) > > I was wondering why this SDK wasn't already part of the SuSe or FC7 > releases. Ten minutes of research later, the answer is: it's proprietary. Part of it is proprietary, most of it is not. The part we are likely to use is all distributed under the GPL, LGPL or IBM Common License. This includes binutils, the two versions of gcc (one for the PPE, one for the SPE), gdb, libspe2 (the code that provides the bridge between the Linux user-space code and the SPEs), etc. > Why would anyone on this mailing list want to install proprietary > compilers? The IBM license agreement is extortionate. Besides the > usual rape and pillage, it says that: > > > http://www14.software.ibm.com/cgi-bin/weblap/lap.pl?la_formnum=&li_formnum=L-MCHN-6MVMPV&title=XL%20C/C%2B%2B%20Alpha%20Edition%20for%20Cell%20Processor > > * You are not authorized to use the Program for productive purposes > * THE PROGRAM MAY CONTAIN A DISABLING DEVICE THAT WILL PREVENT IT FROM > BEING USED AFTER THE EVALUATION PERIOD ENDS. YOU MAY NOT TAMPER WITH THIS > DISABLING DEVICE OR THE PROGRAM. YOU SHOULD TAKE PRECAUTIONS TO AVOID ANY > LOSS OF DATA THAT MIGHT RESULT WHEN THE PROGRAM CAN NO LONGER BE USED. > * You assign to IBM all right, title, and interest (including ownership of > copyright) in any data, suggestions, or written > materials that 1) are related to the Program and 2) You provide to IBM. > * Your right to run the program "ends after 90 days." We're not using those compilers. We're not recommending them. I don't even install them. And I get your point ;) > If the above wasn't enough, it's also against GNU project policy to > use the mailing list or project documentation to advertise or advocate > for proprietary software. > > If the GNU compilers for the CELL aren't good enough for us, I suggest > that we improve them. If we just can't exploit the CELL processor without > proprietary software and patented algorithms, then I suggest we go > back to focusing on hosting GNU Radio on hardware that comes with freedom. If you've got some time, please put together a script, etc that pulls the free stuff down from the Barcelona Supercomputer Center. That's where it's all hosted. Then we can point users at that script and not at the IBM SDK site. Here's a good starting point: http://www.bsc.es/plantillaH.php?cat_id=431 FYI, the kernel and oprofile they're distributing only work on bare metal (e.g., IBM QS20 or QS21 bladeserver), not on the hypervisored PS3. There's currently lots of churn on the Cell kernels. You can watch the sausage being made on the cbe-oss-dev mailing list. The IBM and Sony Linux hackers are there. https://ozlabs.org/mailman/listinfo/cbe-oss-dev Thanks, Eric ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 versus freedom
> If you want to run on the PS3, you're most likely going to want the > IBM SDK 3.0. The SDK really, really wants FC 7 on the PS3. Some people will do anything for crunchons -- or "promised future" crunchons. (A crunchon is a unit of number-crunching.) I was wondering why this SDK wasn't already part of the SuSe or FC7 releases. Ten minutes of research later, the answer is: it's proprietary. Why would anyone on this mailing list want to install proprietary compilers? The IBM license agreement is extortionate. Besides the usual rape and pillage, it says that: http://www14.software.ibm.com/cgi-bin/weblap/lap.pl?la_formnum=&li_formnum=L-MCHN-6MVMPV&title=XL%20C/C%2B%2B%20Alpha%20Edition%20for%20Cell%20Processor * You are not authorized to use the Program for productive purposes * THE PROGRAM MAY CONTAIN A DISABLING DEVICE THAT WILL PREVENT IT FROM BEING USED AFTER THE EVALUATION PERIOD ENDS. YOU MAY NOT TAMPER WITH THIS DISABLING DEVICE OR THE PROGRAM. YOU SHOULD TAKE PRECAUTIONS TO AVOID ANY LOSS OF DATA THAT MIGHT RESULT WHEN THE PROGRAM CAN NO LONGER BE USED. * You assign to IBM all right, title, and interest (including ownership of copyright) in any data, suggestions, or written materials that 1) are related to the Program and 2) You provide to IBM. * Your right to run the program "ends after 90 days." [I hope Eric hasn't provided IBM a copy of the GNU Radio code that compiles under this compiler: IBM will end up owning the copyright on GNU Radio. This may occur even if Eric just posts the code in a public place where IBM can download it.] If the above wasn't enough, it's also against GNU project policy to use the mailing list or project documentation to advertise or advocate for proprietary software. If the GNU compilers for the CELL aren't good enough for us, I suggest that we improve them. If we just can't exploit the CELL processor without proprietary software and patented algorithms, then I suggest we go back to focusing on hosting GNU Radio on hardware that comes with freedom. John ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 Success at last
On Mon, Oct 08, 2007 at 06:44:38PM -0700, Eric Blossom wrote: > On Mon, Oct 08, 2007 at 03:42:18AM -0400, Robert McGwier wrote: > > > > I have 2.6.23-rc8 running and alsa-audio working nicely except directly > > in gnuradio. One needs to add the PS3.conf to /etc/alsa as instructed > > by Lavan on his Cell addons disk and then configure the sound card and > > the audio test works. We have some little gotcha in the alsa sink that > > I have been unable to find, so I gave up for now. > > After enabling a bit of diagnostic output, it turns out that the PS3 > alsa interface won't accept RW_INTERLEAVED stereo samples. This > shouldn't be hard to work around. We'll need to support the > RW_NONINTERLEAVED or MMAP_NONINTERLEAVED access type in addition to > RW_INTERLEAVED. > > Volunteers? FYI, I opened ticket:190 on this. Eric ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 Success at last
On Mon, Oct 08, 2007 at 03:42:18AM -0400, Robert McGwier wrote: > > I have 2.6.23-rc8 running and alsa-audio working nicely except directly > in gnuradio. One needs to add the PS3.conf to /etc/alsa as instructed > by Lavan on his Cell addons disk and then configure the sound card and > the audio test works. We have some little gotcha in the alsa sink that > I have been unable to find, so I gave up for now. After enabling a bit of diagnostic output, it turns out that the PS3 alsa interface won't accept RW_INTERLEAVED stereo samples. This shouldn't be hard to work around. We'll need to support the RW_NONINTERLEAVED or MMAP_NONINTERLEAVED access type in addition to RW_INTERLEAVED. Volunteers? [EMAIL PROTECTED] audio]$ ./dial_tone.py audio: using audio_alsa PCM name: hw:0,0 Access types: MMAP_INTERLEAVED NO MMAP_NONINTERLEAVED YES MMAP_COMPLEX NO RW_INTERLEAVED NO RW_NONINTERLEAVEDYES Formats: S16_BE YES S24_BE YES Number of channels min channels: 2 max channels: 2 2 channels YES Sample Rates: min rate: 44100 (dir = 0) max rate: 96000 (dir = 0) 8000 NO 16000 NO 22050 NO 32000 NO 44100 YES 48000 YES 96000 YES 192000 NO audio_alsa_sink[hw:0,0]: failed to set access mask: Invalid argument Traceback (most recent call last): File "./dial_tone.py", line 55, in my_top_block().run() File "./dial_tone.py", line 48, in __init__ dst = audio.sink (sample_rate, options.audio_output) File "/usr/local/lib/python2.5/site-packages/gnuradio/audio_alsa.py", line 236, in sink return _audio_alsa.sink(*args) RuntimeError: audio_alsa_sink > I had hoped when I introduced portaudio that it would be useful to > windows and some of the "other" PC's. It has again. I configured > portaudio to only support alsa and jack. I modified a few python audio > scripts to use the audio_portaudio sink and they work perfectly. You can avoid modifying code by specifying the default audio module in ~/.gnuradio/config.conf E.g., [audio] verbose = True #audio_module = audio_portaudio #audio_module = audio_oss audio_module = audio_alsa > I am not cross compiling yet so it only took 3 hours for gnuradio to > compile from scratch (!!) on the PS3 but this is enough for me to plug > up the usrp later today and see what works. After I turned off all unnecessary services, and changed the initdefault from 5 to 3 in /etc/inittab, it took under 1 hour ;) FYI, even when running in init mode 3 (no X login), you can still start an X session from the text mode command line using $ startx Eric ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
[Discuss-gnuradio] PS3 Success at last
A month or so ago Eric found a compiler bug that prevented swig and python from playing together well on the PS3. As Eric has told you, we have been working on the kernel and Eric almost has the install down pat and documented sufficiently well to tell everyone how. I have been steadily beating on this since my move to the University of Maryland campus as I need it for my projects. I have 2.6.23-rc8 running and alsa-audio working nicely except directly in gnuradio. One needs to add the PS3.conf to /etc/alsa as instructed by Lavan on his Cell addons disk and then configure the sound card and the audio test works. We have some little gotcha in the alsa sink that I have been unable to find, so I gave up for now. I had hoped when I introduced portaudio that it would be useful to windows and some of the "other" PC's. It has again. I configured portaudio to only support alsa and jack. I modified a few python audio scripts to use the audio_portaudio sink and they work perfectly. I am not cross compiling yet so it only took 3 hours for gnuradio to compile from scratch (!!) on the PS3 but this is enough for me to plug up the usrp later today and see what works. I can see that we will be able to take off and work hard on the PS3 now. A suggestion for those who are building and installing Lavan's kernel. The PS3 runs faster and the footprint is noticably smaller if you disable kernel hacking/debug in make menuconfig. Lavan's ps3_defconfig is for kernel hackers as they try to debug the faults. The latest has a fault in that we are back to ps3-boot-game-os and reboot not completing resulting in the need to force power off/on to reboot but this was sorted out before and worked nicely. This is a work in progress but I can definitely see work is ready to begin in earnest. Bob -- AMSAT Director and VP Engineering. Member: ARRL, AMSAT-DL, TAPR, Packrats, NJQRP, QRP ARCI, QCWA, FRC. ARRL SDR WG Chair “An optimist may see a light where there is none, but why must the pessimist always run to blow it out?” Descartes ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 and OpenSuSE 10.3
On Sat, Oct 06, 2007 at 04:38:20PM -0400, Eric A. Cottrell wrote: > Hello, > > I noticed today that OpenSuSE 10.3 is just released so I browsed > around. I will be upgrading to 10.3 in a few days after download. > > I am interested in the PS3 as well so I thought it was great that it is > directly PS3 installable. > http://en.opensuse.org/PS3 > > I remember some comments that Python appears to be broken on the PS3 and > wonder if the Version 2.5.1 packages in 10.3 would address some of the > problems? > > 73 Eric (insert letter after B here) Hi Eric, If you want to run on the PS3, you're most likely going to want the IBM SDK 3.0. The SDK really, really wants FC 7 on the PS3. Although SuSE 10.1 may install, I don't think it's what you want. I'll write up the seriously convoluted install procedure that is required to get FC 7 on the PS3 along with a known-good PS3 specific kernel, and the IBM SDK 3.0 a bit later today. If you want to get a head start, download the Fedora-7-ppc DVD iso and burn it to a DVD: http://torrent.fedoraproject.org/ You'll also need Geoff Levand's (Sony Linux kernel guy) 2007-08-31 ADDON disk: http://www.kernel.org/pub/linux/kernel/people/geoff/cell/CELL-Linux-CL_20070831-ADDON.iso Download and burn it to CD. It's got a PS3 specific 2.6.23 kernel on it, along with the PS3 boot loader, and other magic pieces. Eric, K7GNU ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
[Discuss-gnuradio] PS3 and OpenSuSE 10.3
Hello, I noticed today that OpenSuSE 10.3 is just released so I browsed around. I will be upgrading to 10.3 in a few days after download. I am interested in the PS3 as well so I thought it was great that it is directly PS3 installable. http://en.opensuse.org/PS3 I remember some comments that Python appears to be broken on the PS3 and wonder if the Version 2.5.1 packages in 10.3 would address some of the problems? 73 Eric (insert letter after B here) ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
PS3 audio solution (wasRe: [Discuss-gnuradio] PS3 and gnuradio)
High Performance SDR ( http://hpsdr.org ) has been working on a high end audio interface for SDR (up to 192 kHz bandwidth at baseband, I/Q delivery over USB 2.0). It has a PWM modulator, and even a microphone hookup and codec for microphone and speakers. The set of boards required is Atlas (because it holds up all of the experimental boards), Janus (because he was god of doorways, beginnings and endings), and Ozy (short for Ozymandias, the Shelley "ode"-ed king of old). The Atlas backplane/bus is REALLY expensive. I don't know if most can afford its $28 price tag. Janus is $179 and Ozy is $149. That is cheaper than a really high end sound card. The Janus uses the AKM5394A, the best audio A/D around so far as I can tell. It is used in the Lynx Studios sound cards, the creative Emu, and in the Flex Radio 5000 series, all with good reason. (see http://www.tapr.org Beginning late this summer, but not before, after we get GnuRadio to actually run on the PS3, I will work with Frank and others to make this work. It uses libusb and I do not foresee major difficulties. I think we need intially to do a user space interface and then have an alsa plugin to access it. Thanks to Frank for pointing out the obvious. If anyone else wants to tackle this, jump all over it. Bob Jens Osterkamp wrote: Matthias Klose is currently working on ppe/spe toolchain packages for Ubuntu : http://ozlabs.org/pipermail/cbe-oss-dev/2007-June/002385.html Gruß, Jens ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio -- Robert W. McGwier, Ph.D. Center for Communications Research 805 Bunn Drive Princeton, NJ 08540 (609)-924-4600 (sig required by employer) ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 and gnuradio
On Friday 13 July 2007, Tom Rondeau wrote: > Hector Oron wrote: > > Hello, > > > >> Now, I say this, but frankly, I don't like Fedora. If I could use > >> Debian/Ubuntu on the PS3, I would, but for the tools support, this is > >> just the way to go. > > > > Just in case you did not know: > > > > Installing Cell BE SDK V2.0 on a Debian Host > > http://blog.perlplexity.org/?p=3 > > > > For installing instructions for Debian on Ps3 > > http://www.keshi.org/moin/moin.cgi/PS3/Debian/Live > > > > Ubuntu's one > > http://ubuntuforums.org/showthread.php?t=316047 > > > > Regards > > I had not seen the first one on getting the SDK onto Debian. The one I > saw used alien to convert to deb files, but that didn't work. I'll have > to check this out soon. Matthias Klose is currently working on ppe/spe toolchain packages for Ubuntu : http://ozlabs.org/pipermail/cbe-oss-dev/2007-June/002385.html Gruß, Jens ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 and gnuradio
Hector Oron wrote: > Hello, > >> Now, I say this, but frankly, I don't like Fedora. If I could use >> Debian/Ubuntu on the PS3, I would, but for the tools support, this is >> just the way to go. > > Just in case you did not know: > > Installing Cell BE SDK V2.0 on a Debian Host > http://blog.perlplexity.org/?p=3 > > For installing instructions for Debian on Ps3 > http://www.keshi.org/moin/moin.cgi/PS3/Debian/Live > > Ubuntu's one > http://ubuntuforums.org/showthread.php?t=316047 > > Regards I had not seen the first one on getting the SDK onto Debian. The one I saw used alien to convert to deb files, but that didn't work. I'll have to check this out soon. Thanks for the link! Tom ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 and gnuradio
That needs a Linux ALSA sink. Bob ' Johnathan Corgan wrote: Philip Balister wrote: What would be a good GNU Radio test on a 500 MHz PPC board, with a USRP, but no display? To just test that GNU Radio itself is sane, you can run the dial tone demo in: gnuradio-examples/python/audio/dial_tone.py This will send a dial tone to the audio output. You can test the USB <--> USRP performance using: gnuradio-examples/python/usrp/benchmark_usb.py If you have a suitable receiver board, you can test the whole system by receiving a broadcast FM signal: gnuradio-examples/python/usrp/usrp_wfm_rcv_nogui.py (You need to set which receive side with -R and frequency with -f) -- AMSAT Director and VP Engineering. Member: ARRL, AMSAT-DL, TAPR, Packrats, NJQRP, QRP ARCI, QCWA, FRC. ARRL SDR WG Chair "If you're going to be crazy, you have to get paid for it or else you're going to be locked up." Hunter S. Thompson ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 and gnuradio
Philip Balister wrote: > What would be a good GNU Radio test on a 500 MHz PPC board, with a > USRP, but no display? To just test that GNU Radio itself is sane, you can run the dial tone demo in: gnuradio-examples/python/audio/dial_tone.py This will send a dial tone to the audio output. You can test the USB <--> USRP performance using: gnuradio-examples/python/usrp/benchmark_usb.py If you have a suitable receiver board, you can test the whole system by receiving a broadcast FM signal: gnuradio-examples/python/usrp/usrp_wfm_rcv_nogui.py (You need to set which receive side with -R and frequency with -f) -- Johnathan Corgan Corgan Enterprises LLC http://corganenterprises.com ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 and gnuradio
OK, using OpenEmbedded, I built a file sysem image with GNU Radio (from svn) for my EFIKA board. usrper can set the led fine :) Now, I am not a GNU Radio user. I do have these files in /usr/bin [EMAIL PROTECTED]:~$ usrp usrp_cal_dc_offset usrp_ra_receiver.pyusrper usrp_flex.py usrp_radar_mono.py usrp_psr_receiver.py usrp_sounder.py What would be a good GNU Radio test on a 500 MHz PPC board, with a USRP, but no display? Philip On 7/7/07, Robert McGwier <[EMAIL PROTECTED]> wrote: So far I have been unable to get the python to work properly on any Power PC running Linux. It runs on Mac OSX but not PPC. For example, make check fails in a segmentation error in the base routine that runs the actual checks. It is no trouble at all to get GnuRadio to make and install on the PS3 running FC6. It took me about 3 hours. It is just that nothing runs after that. Eric Blossom is going to begin work on getting GnuRadio to run on the Cell processor after he returns from vacation later this month. I am certain he will resolve this issue early in the process. Bob Clark Pope wrote: > > I see that you can buy a PS3 now with YDL linux installed. > (http://www.terrasoftsolutions.com/products/sony/) How are the efforts > to get gnu radio running on it going? > > I don't play video games but I'm seriously considering buying to get the > bluray player, Would be great if I can run gnuradio too. > > Thanks, > Clark > > _ > http://liveearth.msn.com > > > > ___ > Discuss-gnuradio mailing list > Discuss-gnuradio@gnu.org > http://lists.gnu.org/mailman/listinfo/discuss-gnuradio > -- AMSAT Director and VP Engineering. Member: ARRL, AMSAT-DL, TAPR, Packrats, NJQRP, QRP ARCI, QCWA, FRC. ARRL SDR WG Chair "If you're going to be crazy, you have to get paid for it or else you're going to be locked up." Hunter S. Thompson ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 and gnuradio
Hello, Now, I say this, but frankly, I don't like Fedora. If I could use Debian/Ubuntu on the PS3, I would, but for the tools support, this is just the way to go. Just in case you did not know: Installing Cell BE SDK V2.0 on a Debian Host http://blog.perlplexity.org/?p=3 For installing instructions for Debian on Ps3 http://www.keshi.org/moin/moin.cgi/PS3/Debian/Live Ubuntu's one http://ubuntuforums.org/showthread.php?t=316047 Regards -- Héctor Orón ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 and gnuradio
Luke Scharf wrote: > Clark Pope wrote: >> Anyway, you're saying the way to go is FC6 and not YDL, right? Maybe >> Eric can confirm he plans to use FC6? > > I've been on the YDL list for a while, since we use them for their > Apple XServe support. (I run FC6 on my G5 Tower desktop, for the > reasons that I'll mention shortly). > > From the discussions I've read, it seems that if you want to fully > exploit the Cell parts of the CPU, YDL is probably the way to go -- at > least they're packaging a lot of Cell-specific development tools that > look like they could be useful. Also, the TerraSoft crew is quite > responsive to the needs of their scientific/numerical users. > > On the other hand, if you want something that looks like a regular > Linux desktop, then Fedora is certainly the way to go -- there are a > lot more pre-compiled packages, especially WRT desktop and multimedia > applications. Can anyone say anything more about Cell support on Fedora? > > Why not try both and see which one matches your needs better? > > -Luke Most of IBM's development tools are built for FC5/6. I tried working with them under Ubuntu, but the amount of time to get everything working wasn't worth it and they worked out of the box/rpm under an FC6 installation. This stuff includes the SPU/PPU GCC compiler's, IBM's CELL compilers, the libspe2 library, the binutils, etc., as well as their simulation environment for not running on a cell machine. Now, I say this, but frankly, I don't like Fedora. If I could use Debian/Ubuntu on the PS3, I would, but for the tools support, this is just the way to go. Tom ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 and gnuradio
From: Luke Scharf <[EMAIL PROTECTED]> To: Clark Pope <[EMAIL PROTECTED]> CC: [EMAIL PROTECTED], discuss-gnuradio@gnu.org Subject: Re: [Discuss-gnuradio] PS3 and gnuradio Date: Thu, 12 Jul 2007 10:56:22 -0400 Clark Pope wrote: Anyway, you're saying the way to go is FC6 and not YDL, right? Maybe Eric can confirm he plans to use FC6? I've been on the YDL list for a while, since we use them for their Apple XServe support. (I run FC6 on my G5 Tower desktop, for the reasons that I'll mention shortly). From the discussions I've read, it seems that if you want to fully exploit the Cell parts of the CPU, YDL is probably the way to go -- at least they're packaging a lot of Cell-specific development tools that look like they could be useful. Also, the TerraSoft crew is quite responsive to the needs of their scientific/numerical users. On the other hand, if you want something that looks like a regular Linux desktop, then Fedora is certainly the way to go -- there are a lot more pre-compiled packages, especially WRT desktop and multimedia applications. Can anyone say anything more about Cell support on Fedora? Why not try both and see which one matches your needs better? -Luke My main interest is signal processing (though I'm thinking about playing with mythtv), so I'm just looking for the the load that is most likely to be supported by gnuradio. I guess I'll go with whatever eric picks when he gets back. -Clark _ Local listings, incredible imagery, and driving directions - all in one place! http://maps.live.com/?wip=69&FORM=MGAC01 ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 and gnuradio
Clark Pope wrote: Anyway, you're saying the way to go is FC6 and not YDL, right? Maybe Eric can confirm he plans to use FC6? I've been on the YDL list for a while, since we use them for their Apple XServe support. (I run FC6 on my G5 Tower desktop, for the reasons that I'll mention shortly). From the discussions I've read, it seems that if you want to fully exploit the Cell parts of the CPU, YDL is probably the way to go -- at least they're packaging a lot of Cell-specific development tools that look like they could be useful. Also, the TerraSoft crew is quite responsive to the needs of their scientific/numerical users. On the other hand, if you want something that looks like a regular Linux desktop, then Fedora is certainly the way to go -- there are a lot more pre-compiled packages, especially WRT desktop and multimedia applications. Can anyone say anything more about Cell support on Fedora? Why not try both and see which one matches your needs better? -Luke smime.p7s Description: S/MIME Cryptographic Signature ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 and gnuradio
Eric might still be on vacation in South America or Europe! Eric is running triple boot, FC5, FC6, and YDL. I run FC6 and GnuRadio compiles just fine but does not run. I have asked here if there is ANYONE who has GnuRadio compiling and running on ANY kind of power PC running FC6 and received no answer. Work on porting relevant routines from GnuRadio to the PS3 is being supported by contract. I am expecting that work to begin in earnest this month. The bureaucrats are in the way of rapid progress but we now it will happen. Bob Clark Pope wrote: Well after the $100 price drop I couldn't resist so I have my 60 GB PS3 now. Bought a used package so probably only be ~$350 out of pocket after I sell off the games. Probably half what a stand alone bluray player will cost you. Didn't realize it was designed to host an extra OS. I thought people were hacking it to get Linux on. Turns out you can have Linux AND all the existing video game console functionality. Anyway, you're saying the way to go is FC6 and not YDL, right? Maybe Eric can confirm he plans to use FC6? Thanks, Clark -- Robert W. McGwier, Ph.D. Center for Communications Research 805 Bunn Drive Princeton, NJ 08540 (609)-924-4600 (sig required by employer) ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 and gnuradio
Well after the $100 price drop I couldn't resist so I have my 60 GB PS3 now. Bought a used package so probably only be ~$350 out of pocket after I sell off the games. Probably half what a stand alone bluray player will cost you. Didn't realize it was designed to host an extra OS. I thought people were hacking it to get Linux on. Turns out you can have Linux AND all the existing video game console functionality. Anyway, you're saying the way to go is FC6 and not YDL, right? Maybe Eric can confirm he plans to use FC6? Thanks, Clark From: Tom Rondeau <[EMAIL PROTECTED]> To: GNU Radio Discuss Subject: Re: [Discuss-gnuradio] PS3 and gnuradio Date: Sat, 07 Jul 2007 22:34:56 +0100 You can run GNU Radio using the new version of the hierarchical blocks. The example in gnuradio-examples/python/hier/usrp/usrp_siggen.py runs on PS3 as is. The reason is due to more of the scheduling being handled in C++ and very minimal work in Python. I tracked down the problem in the Python scheduler where, for some reason, the memory addresses of the blocks get corrupted. I didn't go any further to figure out why, though. Of course, it just runs on the PPE element, and there is still plenty of work to be done to make real use of the CELL processor properly. Tom Robert McGwier wrote: So far I have been unable to get the python to work properly on any Power PC running Linux. It runs on Mac OSX but not PPC. For example, make check fails in a segmentation error in the base routine that runs the actual checks. It is no trouble at all to get GnuRadio to make and install on the PS3 running FC6. It took me about 3 hours. It is just that nothing runs after that. Eric Blossom is going to begin work on getting GnuRadio to run on the Cell processor after he returns from vacation later this month. I am certain he will resolve this issue early in the process. Bob Clark Pope wrote: I see that you can buy a PS3 now with YDL linux installed. (http://www.terrasoftsolutions.com/products/sony/) How are the efforts to get gnu radio running on it going? I don't play video games but I'm seriously considering buying to get the bluray player, Would be great if I can run gnuradio too. Thanks, Clark ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio _ http://newlivehotmail.com ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 and gnuradio
You can run GNU Radio using the new version of the hierarchical blocks. The example in gnuradio-examples/python/hier/usrp/usrp_siggen.py runs on PS3 as is. The reason is due to more of the scheduling being handled in C++ and very minimal work in Python. I tracked down the problem in the Python scheduler where, for some reason, the memory addresses of the blocks get corrupted. I didn't go any further to figure out why, though. Of course, it just runs on the PPE element, and there is still plenty of work to be done to make real use of the CELL processor properly. Tom Robert McGwier wrote: So far I have been unable to get the python to work properly on any Power PC running Linux. It runs on Mac OSX but not PPC. For example, make check fails in a segmentation error in the base routine that runs the actual checks. It is no trouble at all to get GnuRadio to make and install on the PS3 running FC6. It took me about 3 hours. It is just that nothing runs after that. Eric Blossom is going to begin work on getting GnuRadio to run on the Cell processor after he returns from vacation later this month. I am certain he will resolve this issue early in the process. Bob Clark Pope wrote: I see that you can buy a PS3 now with YDL linux installed. (http://www.terrasoftsolutions.com/products/sony/) How are the efforts to get gnu radio running on it going? I don't play video games but I'm seriously considering buying to get the bluray player, Would be great if I can run gnuradio too. Thanks, Clark ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 and gnuradio
So far I have been unable to get the python to work properly on any Power PC running Linux. It runs on Mac OSX but not PPC. For example, make check fails in a segmentation error in the base routine that runs the actual checks. It is no trouble at all to get GnuRadio to make and install on the PS3 running FC6. It took me about 3 hours. It is just that nothing runs after that. Eric Blossom is going to begin work on getting GnuRadio to run on the Cell processor after he returns from vacation later this month. I am certain he will resolve this issue early in the process. Bob Clark Pope wrote: I see that you can buy a PS3 now with YDL linux installed. (http://www.terrasoftsolutions.com/products/sony/) How are the efforts to get gnu radio running on it going? I don't play video games but I'm seriously considering buying to get the bluray player, Would be great if I can run gnuradio too. Thanks, Clark _ http://liveearth.msn.com ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio -- AMSAT Director and VP Engineering. Member: ARRL, AMSAT-DL, TAPR, Packrats, NJQRP, QRP ARCI, QCWA, FRC. ARRL SDR WG Chair "If you're going to be crazy, you have to get paid for it or else you're going to be locked up." Hunter S. Thompson ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
[Discuss-gnuradio] PS3 and gnuradio
I see that you can buy a PS3 now with YDL linux installed. (http://www.terrasoftsolutions.com/products/sony/) How are the efforts to get gnu radio running on it going? I don't play video games but I'm seriously considering buying to get the bluray player, Would be great if I can run gnuradio too. Thanks, Clark _ http://liveearth.msn.com ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
[Discuss-gnuradio] PS3 and dial_tone
With help from Jonathan Corgan (getting a file fixed), I was able to make the executable gnuradio-example/c++/dial_done/dial_tone on the PS3 under Yellow Dog Linux. It failed setting the access mask and that was when I realize that the PS3 has one of these monster devices with all of these strange outputs so I added the argument plughw:0,0 to the sink make request. It worked perfectly. This is a simple example, but it does exercise a healthy bunch of the core. We need to get the usrp brought into the c++ world as right now it can be connected to a processing graph primarily (only ?) with python. Bob -- AMSAT Director and VP Engineering. Member: ARRL, AMSAT-DL, TAPR, Packrats, NJQRP, QRP ARCI, QCWA, FRC. ARRL SDR WG Chair "Taking fun as simply fun and earnestness in earnest shows how thoroughly thou none of the two discernest." - Piet Hine ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 success with audio
Eric Blossom wrote: On Thu, Mar 08, 2007 at 08:15:47PM -0500, Robert McGwier wrote: We definitely have work to do. Portaudio is the ONLY thing I can get to work the PS3 sound system but not in gnuradio. The pieces in the bin directory for testing portaudio that open the sink only work. I can produce tones, hundreds of tones, do latency tests, etc. gr-audio-XXX fails in all cases, including gr-audio-portaudio. I do not believe this is a python issue but I am unsure. The error message is a bit strange. If you download the svn for portaudio, and do configure; make; sudo make install, it puts the portaudio-2.0.pc in /usr/local/lib/pkgconfig unless you specify another prefix. Portaudio is working on the ps3 and the files in bin that begin patest_(produce a sinusoid of some type) make tones and work. Bob Given that the Python stuff is still broken on PS3, I'm not surprised than gr-audio-* doesn't work. Just let me say that I am not certain this is a python error but until make check works, I cannot easily rule it out. Portaudio definitely works on the PS3 so we will have audio when we get stuff rolling. Eric Thanks, Bob -- AMSAT Director and VP Engineering. Member: ARRL, AMSAT-DL, TAPR, Packrats, NJQRP, QRP ARCI, QCWA, FRC. ARRL SDR WG Chair "Taking fun as simply fun and earnestness in earnest shows how thoroughly thou none of the two discernest." - Piet Hine ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 success with audio
On Thu, Mar 08, 2007 at 08:15:47PM -0500, Robert McGwier wrote: > We definitely have work to do. Portaudio is the ONLY thing I can get to > work the PS3 sound system but not in gnuradio. > > The pieces in the bin directory for testing portaudio that open the sink > only work. I can produce tones, hundreds of tones, do latency tests, etc. > > gr-audio-XXX fails in all cases, including gr-audio-portaudio. > > I do not believe this is a python issue but I am unsure. The error > message is a bit strange. > > If you download the svn for portaudio, and do configure; make; sudo make > install, it puts the portaudio-2.0.pc in /usr/local/lib/pkgconfig > unless you specify another prefix. Portaudio is working on the ps3 and > the files in bin that begin patest_(produce a sinusoid of some type) > make tones and work. > > Bob Given that the Python stuff is still broken on PS3, I'm not surprised than gr-audio-* doesn't work. Eric ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 success with audio
We definitely have work to do. Portaudio is the ONLY thing I can get to work the PS3 sound system but not in gnuradio. The pieces in the bin directory for testing portaudio that open the sink only work. I can produce tones, hundreds of tones, do latency tests, etc. gr-audio-XXX fails in all cases, including gr-audio-portaudio. I do not believe this is a python issue but I am unsure. The error message is a bit strange. If you download the svn for portaudio, and do configure; make; sudo make install, it puts the portaudio-2.0.pc in /usr/local/lib/pkgconfig unless you specify another prefix. Portaudio is working on the ps3 and the files in bin that begin patest_(produce a sinusoid of some type) make tones and work. Bob Eric Blossom wrote: On Thu, Mar 08, 2007 at 11:55:05AM -0500, Tom Rondeau wrote: -Original Message- From: [EMAIL PROTECTED] [mailto:discuss- I literally have absolutely no idea what I have done, BUT, snd_ps3 is now found by PS3 running Yellow Dog Linux. At some time in the past two weeks while I was running around, I did an update (after Terra finally got my password fixed) and the sound system is found. I compiled and installed PortAudio and pa_devs finds the device and all of the test programs that are using only output work perfectly. Bob, are you using gr-audio-portaudio? I've installed portaudio from yum on FC6, but it's only on version 18 while the GR configure script is trying to find version 19. On top of that, when I installed version 19, there is no portaudio.pc file for pkg-config to look at. I'm sure this is all due to ignorance on my part with using portaudio devices. Thanks, Tom Tom, I'm not sure, but I'm pretty sure he's using ALSA. Eric ___ -- AMSAT Director and VP Engineering. Member: ARRL, AMSAT-DL, TAPR, Packrats, NJQRP, QRP ARCI, QCWA, FRC. ARRL SDR WG Chair "Taking fun as simply fun and earnestness in earnest shows how thoroughly thou none of the two discernest." - Piet Hine ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3 success with audio
On Thu, Mar 08, 2007 at 11:55:05AM -0500, Tom Rondeau wrote: > > -Original Message- > > From: [EMAIL PROTECTED] [mailto:discuss- > > > > I literally have absolutely no idea what I have done, BUT, snd_ps3 > > is now found by PS3 running Yellow Dog Linux. At some time in the > > past two weeks while I was running around, I did an update (after Terra > > finally got my password fixed) and the sound system is found. I compiled > > and installed PortAudio and pa_devs finds the device and all of the test > > programs that are using only output work perfectly. > > Bob, are you using gr-audio-portaudio? I've installed portaudio from yum on > FC6, but it's only on version 18 while the GR configure script is trying to > find version 19. On top of that, when I installed version 19, there is no > portaudio.pc file for pkg-config to look at. > > I'm sure this is all due to ignorance on my part with using portaudio > devices. > > Thanks, > Tom Tom, I'm not sure, but I'm pretty sure he's using ALSA. Eric ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
RE: [Discuss-gnuradio] PS3 success with audio
> -Original Message- > From: [EMAIL PROTECTED] [mailto:discuss- > > I literally have absolutely no idea what I have done, BUT, snd_ps3 > is now found by PS3 running Yellow Dog Linux. At some time in the > past two weeks while I was running around, I did an update (after Terra > finally got my password fixed) and the sound system is found. I compiled > and installed PortAudio and pa_devs finds the device and all of the test > programs that are using only output work perfectly. Bob, are you using gr-audio-portaudio? I've installed portaudio from yum on FC6, but it's only on version 18 while the GR configure script is trying to find version 19. On top of that, when I installed version 19, there is no portaudio.pc file for pkg-config to look at. I'm sure this is all due to ignorance on my part with using portaudio devices. Thanks, Tom ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
[Discuss-gnuradio] PS3 success with audio
I literally have absolutely no idea what I have done, BUT, snd_ps3 is now found by PS3 running Yellow Dog Linux. At some time in the past two weeks while I was running around, I did an update (after Terra finally got my password fixed) and the sound system is found. I compiled and installed PortAudio and pa_devs finds the device and all of the test programs that are using only output work perfectly. When you run "top" on this machine, you will see two threads. Do NOT do make -j3 or -j2. Swig output has been compiling for AN HOUR and has not finished. ;-). For those running YDL, I am sure you have discovered that you have about 8k of free space in RAM when running Enlightenment. Gnome is even worse (2k). In /etc/inittab where you find id:5:initdefault change the 5 to a 3 and reboot to get rid of the X login. If you need X things, ssh in from a remote terminal. Now you will find that about 50% of RAM is free when the machine is sitting idle. The drop in swapping has a major impact on the PS3. I am hoping (like Tom) to report some success with the PS3 on Sunday. I have a full plate with Frank Brickle and Flex until Sunday afternoon. Bob -- AMSAT Director and VP Engineering. Member: ARRL, AMSAT-DL, TAPR, Packrats, NJQRP, QRP ARCI, QCWA, FRC. ARRL SDR WG Chair "Taking fun as simply fun and earnestness in earnest shows how thoroughly thou none of the two discernest." - Piet Hine ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
[Discuss-gnuradio] PS3 one more time
http://n4hy.org/ps3_fedora.jpg http://n4hy.org/ps3_dttsp.jpg http://n4hy.org/ps3_gnuradio.jpg Now to accomplish something useful. Bob -- AMSAT Director and VP Engineering. Member: ARRL, AMSAT-DL, TAPR, Packrats, NJQRP, QRP ARCI, QCWA, FRC. ARRL SDR WG Chair "Taking fun as simply fun and earnestness in earnest shows how thoroughly thou none of the two discernest." - Piet Hine ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
[Discuss-gnuradio] PS3 shortage over?
My local Target, Walmart, and others have PS3's on the shelf. They have controllers, and accessories galore. That said, even with shipping, I saved money (taxes, etc.) by buying from Fry's. My early purchase from Terra (Yellow Dog) will be consumed by others when that comes to fruition so I didn't even lose my $100. I hope to have Fedora Core 5 on it next Thursday after I return from two days of working on HPSDR things in Maryland and will report my PS3/Linux results here. I assume you will not be interested in my wife's Tiger Woods golf score so I won't be reporting that. ;-) Bob -- AMSAT Director and VP Engineering. Member: ARRL, AMSAT-DL, TAPR, Packrats, NJQRP, QRP ARCI, QCWA, FRC. ARRL SDR WG Chair "Taking fun as simply fun and earnestness in earnest shows how thoroughly thou none of the two discernest." - Piet Hine ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3/Cell BE platform
Newell, You might take a look at http://www.gpgpu.org/ From their web site: > GPGPU stands for General-Purpose computation on GPUs. With the increasing programmability of commodity graphics processing units (GPUs), these chips are capable of performing more than the specific graphics computations for which they were designed. They are now capable coprocessors, and their high speed makes them useful for a variety of applications. The goal of this page is to catalog the current and historical use of GPUs for general-purpose computation. There's a fairly large amount of activity going on in this area, and there's some open source code to look at, e.g., http://sourceforge.net/projects/gpgpu/ At a minimum, you'd think the GPU on a PC graphics card could serve as a "smart sink" capable of doing its own histograms, FFT's, waterfalls, etc. for display purposes. Maybe more... Steve On Nov 19, 2006, at 5:19 PM, Newell Jensen wrote: We'd have to compile the blocks for the SPE's. Hopefully the compiler does a decent job of extracting parallelism from the code without too much help. Of course we could hand tune the filter kernels like we do today. The scheduler and behind-the-scenes block interconnect/buffering would need to be modified, but that's pretty well abstracted away from the user's view of the world. We'd need to build some tools that would allow us to measure performance on the SPE and allow us to do feedback based assignment of subsets of blocks in a graph to the SPEs. I think this would be an iterative process. That is, partition the graph across the SPEs. Run your test case. Measure. Repeat. Sounds like fun. And yes, I think we could get the HDTV receiver running in real time ;) Eric I am looking to build a new computer (or buy) for graduate school and I would like my platform to be able to have real-time image-processing capabilities. Do you know of any other platforms where that is possible right now? Also, what do you think of the possibility of developing another peripheal that would act like a Cell BE board that would be used in conjunction with the USRP? Newell _ View Athletes Collections with Live Search http://sportmaps.live.com/index.html? source=hmemailtaglinenov06&FORM=MGAC01 ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3/Cell BE platform
We'd have to compile the blocks for the SPE's. Hopefully the compiler does a decent job of extracting parallelism from the code without too much help. Of course we could hand tune the filter kernels like we do today. The scheduler and behind-the-scenes block interconnect/buffering would need to be modified, but that's pretty well abstracted away from the user's view of the world. We'd need to build some tools that would allow us to measure performance on the SPE and allow us to do feedback based assignment of subsets of blocks in a graph to the SPEs. I think this would be an iterative process. That is, partition the graph across the SPEs. Run your test case. Measure. Repeat. Sounds like fun. And yes, I think we could get the HDTV receiver running in real time ;) Eric I am looking to build a new computer (or buy) for graduate school and I would like my platform to be able to have real-time image-processing capabilities. Do you know of any other platforms where that is possible right now? Also, what do you think of the possibility of developing another peripheal that would act like a Cell BE board that would be used in conjunction with the USRP? Newell _ View Athletes Collections with Live Search http://sportmaps.live.com/index.html?source=hmemailtaglinenov06&FORM=MGAC01 ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3/Cell BE platform
On Thu, Nov 16, 2006 at 10:10:23PM +, Newell Jensen wrote: > As I am new to all this please correct me if I am wrong as I am trying to > get a better understanding of everything. I am assuming that even though > you are able run linux on the PS3 that you will not be able to use gnuradio > as it stands because of the different processor architechture, right? That > is, the code in gnuradio wouldn't take advantage of the 7 SPEs etc.? With > reading over what I have, it seems like using the Cell BE would be a great > thing and the future looks bright for this. Thanks > > Newell We'd have to compile the blocks for the SPE's. Hopefully the compiler does a decent job of extracting parallelism from the code without too much help. Of course we could hand tune the filter kernels like we do today. The scheduler and behind-the-scenes block interconnect/buffering would need to be modified, but that's pretty well abstracted away from the user's view of the world. We'd need to build some tools that would allow us to measure performance on the SPE and allow us to do feedback based assignment of subsets of blocks in a graph to the SPEs. I think this would be an iterative process. That is, partition the graph across the SPEs. Run your test case. Measure. Repeat. Sounds like fun. And yes, I think we could get the HDTV receiver running in real time ;) Eric ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3/Cell BE platform
On 11/16/06, Jason <[EMAIL PROTECTED]> wrote: Personally, I would only buy a PS3 for SDR if you either a.) have disposable income. b.) play a lot of games, and are willing to save up the cash. And that's only if your interested in porting GnuRadio to the Cell BE. Currently, nothing has been ported, let alone tested on it as is. If you just want to dive into SDR, I would advise against it. I would recommend running simulations using their Cell software suite, and get something working within the simulator before going ahead and buying any hardware. Playing with a Cell BE processor would be interesting, but there are plenty of things you can do within simulation first that can be just as useful without spending $600 on a PS3 (if you can find one this holiday season!). Brian ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3/Cell BE platform
Newell Jensen wrote: Yes and no. From casual observation (I don't have a ppc platform currently), all of gnuradio's dependencies already run on most linux ppc and ppc64 distributions. So getting it up and running shouldn't be too difficult. The fun part will be modding GnuRadio and its dependencies to take advantage of the SPEs. So, even though there is a libfft (for example) running on ppc/ppc64, it doesn't take advantage of the SPEs. To do so is a bit more involved than just swapping out some math instructions/functions. It involves, from preliminary review, wrapping the chunk of code you want on the SPE in an spu_thread, shipping data in and out via DMA, minimizing ooo (out of order, ie branchy) code, then reworking the algorithm to take advantage of the SIMD (single instruction, multiple data) instruction set. Once that's done, recompile it with IBM's modified GNU toolchain, and watch it crash. :) Thanks for the info, it helped. From what I have read it seems like SP3 is the only way to currently get a hold of the Cell BE, without having to buy an expensive workstation or server, right? I am just curious because I am looking to get a new computer soon that I will be using for graduate school working on SDR (As of now, I am looking to work on GPR and possible improvements of this), and from the reading I have done it seems like this would be a great architecture for SDR. Curious if you knew if they were going to offer this chipset for desktops or what your opinion about switching to that platform was? From my digging, the PS3 is the only way for the average joe to get a CellBE right now. It may be good for SDR in the future, but keep in mind, it doesn't have any PCI/PCI-X/whatever slots, the drive is a 60GB 2.5" SATA, the ram is soldered down, and so there is little to no room for upgrades. Which isn't a show stopper, it's just fairly limiting. On the software side, it's going to take a while for the OSS hackers to figure out how to best take advantage of the Cell and mod their packages. So unless you're willing to dive in, it's going to be a while. Personally, I would only buy a PS3 for SDR if you either a.) have disposable income. b.) play a lot of games, and are willing to save up the cash. And that's only if your interested in porting GnuRadio to the Cell BE. Currently, nothing has been ported, let alone tested on it as is. If you just want to dive into SDR, I would advise against it. hth, Jason. ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3/Cell BE platform
On Friday 17 November 2006 09:42, Jason wrote: > instructions/functions. It involves, from preliminary review, wrapping > the chunk of code you want on the SPE in an spu_thread, shipping data in > and out via DMA, minimizing ooo (out of order, ie branchy) code, then > reworking the algorithm to take advantage of the SIMD (single > instruction, multiple data) instruction set. Once that's done, > recompile it with IBM's modified GNU toolchain, and watch it crash. :) I wouldn't be surprised if a patch for fftw turns up pretty quickly.. I imagine the code in there is already fairly branch free and organised for SIMD operation because of the existing MMX/SSE optimisations. -- Daniel O'Connor software and network engineer for Genesis Software - http://www.gsoft.com.au "The nice thing about standards is that there are so many of them to choose from." -- Andrew Tanenbaum GPG Fingerprint - 5596 B766 97C0 0E94 4347 295E E593 DC20 7B3F CE8C pgp5VlODzP4Mk.pgp Description: PGP signature ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3/Cell BE platform
Yes and no. From casual observation (I don't have a ppc platform currently), all of gnuradio's dependencies already run on most linux ppc and ppc64 distributions. So getting it up and running shouldn't be too difficult. The fun part will be modding GnuRadio and its dependencies to take advantage of the SPEs. So, even though there is a libfft (for example) running on ppc/ppc64, it doesn't take advantage of the SPEs. To do so is a bit more involved than just swapping out some math instructions/functions. It involves, from preliminary review, wrapping the chunk of code you want on the SPE in an spu_thread, shipping data in and out via DMA, minimizing ooo (out of order, ie branchy) code, then reworking the algorithm to take advantage of the SIMD (single instruction, multiple data) instruction set. Once that's done, recompile it with IBM's modified GNU toolchain, and watch it crash. :) hth, Jason. Thanks for the info, it helped. From what I have read it seems like SP3 is the only way to currently get a hold of the Cell BE, without having to buy an expensive workstation or server, right? I am just curious because I am looking to get a new computer soon that I will be using for graduate school working on SDR (As of now, I am looking to work on GPR and possible improvements of this), and from the reading I have done it seems like this would be a great architecture for SDR. Curious if you knew if they were going to offer this chipset for desktops or what your opinion about switching to that platform was? Newell _ Talk now to your Hotmail contacts with Windows Live Messenger. http://clk.atdmt.com/MSN/go/msnnkwme002001msn/direct/01/?href=http://get.live.com/messenger/overview ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3/Cell BE platform
Newell Jensen wrote: As I am new to all this please correct me if I am wrong as I am trying to get a better understanding of everything. I am assuming that even though you are able run linux on the PS3 that you will not be able to use gnuradio as it stands because of the different processor architechture, right? That is, the code in gnuradio wouldn't take advantage of the 7 SPEs etc.? With reading over what I have, it seems like using the Cell BE would be a great thing and the future looks bright for this. Thanks Yes and no. From casual observation (I don't have a ppc platform currently), all of gnuradio's dependencies already run on most linux ppc and ppc64 distributions. So getting it up and running shouldn't be too difficult. The fun part will be modding GnuRadio and its dependencies to take advantage of the SPEs. So, even though there is a libfft (for example) running on ppc/ppc64, it doesn't take advantage of the SPEs. To do so is a bit more involved than just swapping out some math instructions/functions. It involves, from preliminary review, wrapping the chunk of code you want on the SPE in an spu_thread, shipping data in and out via DMA, minimizing ooo (out of order, ie branchy) code, then reworking the algorithm to take advantage of the SIMD (single instruction, multiple data) instruction set. Once that's done, recompile it with IBM's modified GNU toolchain, and watch it crash. :) hth, Jason. ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
RE: [Discuss-gnuradio] PS3/Cell BE platform
As I am new to all this please correct me if I am wrong as I am trying to get a better understanding of everything. I am assuming that even though you are able run linux on the PS3 that you will not be able to use gnuradio as it stands because of the different processor architechture, right? That is, the code in gnuradio wouldn't take advantage of the 7 SPEs etc.? With reading over what I have, it seems like using the Cell BE would be a great thing and the future looks bright for this. Thanks Newell _ MSN Shopping has everything on your holiday list. Get expert picks by style, age, and price. Try it! http://shopping.msn.com/content/shp/?ctId=8000,ptnrid=176,ptnrdata=200601&tcode=wlmtagline ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3/Cell BE platform
On 11/16/06, Jason <[EMAIL PROTECTED]> wrote: Legal nitpicks aside, the SPE will apparently take in an array of four single-precision floats (128bits total in one register), and multiply or add it to another array of four floats, all in one instruction. This would massively reduce the number of instructions for, say, multiplying two 4x4 matrices together. Then toss in the fact that there are 7 SPE's, each moving at 3.2GHz, each with 256K of DMA accessible local memory. Sorry, I'm drooling. :) From my understanding, each SPE has 128 registers each with a width of 128 bits. Your program and data that you are currently working on needs to all fit within the local store memory of 256k - so you need to be very aware of where your data is and where it is going. You can then have one SPE dedicated to streaming FFT's to another SPE which is doing some kind of correlation/filtering then followed by whatever else you want the other SPE's to do. It should also be noted that DMA from one SPE to another is not all equal - some combinations work faster than others, but if I remember correctly, you can't choose which SPE you assign programs to - it just chooses one given the interface from IBM. The IBM dev kit has a complete simulator in it that should be able to do everything you might want to do with a Cell, just without the great performance. Brian ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3/Cell BE platform
Robert McGwier wrote: Jason wrote: [snip] has anyone been looking at the Cell BE processor as a gnuradio backend platform? For US$600 you get a Playstation 3 that boots linux, and has 7 128bit RISC processors, each with 256k local memory, all moving at ~3.4GHz. Each RISC processor is optimized for vector math / stream processing... :) There is already an SDK (modified GNU toolchain) from IBM. It seems to use a standard toolchain and apps that run on a 64bit ppc, but the toolchain is modified to compile special threads (different exec type) for the RISC processors (called SPEs). threads are written in C against a provided library. My thought for GnuRadio was to make near realtime mod/demod of ATSC a possibility. As well as speeding up most other stream processing. >> [snip] > Yes. For a work project my group is purchasing two Mercury 1U Cell > servers. This will make ATSC and HDTV a real possibility. These > Mercury servers have a high bandwidth PCI express connection for getting > data in and out of the world.Another fellow and I will be porting > GnuRadio to the Cell BE ont he Mercury. It is not for consumers in > this form. These things are over $10,000 a piece with PPC Linux > installed. I cannot wait until we get the PS3's hacked to be Linux > desktops. Already been done. [1], [2] Hell, it's not even a hack. The PS3 supports installing Yellow Dog out of the box. So your CellBE mods to GnuRadio will be of interest as soon as the Christmas rush is over. :) Jason. [1] http://www.powerdeveloper.org/playstation.php [2] http://ps3.ign.com/articles/739/739688p1.html ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3/Cell BE platform
Thomas Schmid wrote: On 11/15/06, Daniel O'Connor <[EMAIL PROTECTED]> wrote: On Thursday 16 November 2006 16:22, Robert McGwier wrote: > The Nividia GPU's have fft and blas running on them. They are doing > teraflops and the tools/SDK are available under NDA. They do indeed > have multiply , accumulate, etc. ATI/AMD have something similar.. http://ati.amd.com/companyinfo/researcher/resources.html (Not that I have used either) And somebody even did the implementation of the fft on a GPU for gnuradio already: www.ecsl.cs.sunysb.edu/fir/fir.ps Looks kind of old, but very interesting... If you get a copy of the CellSDK iso [1], extract the cell-sdk-1.1 rpm, all source in there (including 'src/lib/fft/*') is under the Common Public License 1.0, which is approved by OSI [2]. The fft code will do 1d and 2d... The only thing which concerns me is that none of the source files in the SDK refer to the license file, they simply say "All rights reserved". Which doesn't exactly give me a warm fuzzy. However, the CPL-1.0 file is in the 'license/' directory just under the root of the SDK tree. It should be reasonable to assume that all files in the SDK fall under that license, but ianal. Legal nitpicks aside, the SPE will apparently take in an array of four single-precision floats (128bits total in one register), and multiply or add it to another array of four floats, all in one instruction. This would massively reduce the number of instructions for, say, multiplying two 4x4 matrices together. Then toss in the fact that there are 7 SPE's, each moving at 3.2GHz, each with 256K of DMA accessible local memory. Sorry, I'm drooling. :) Jason. [1] http://www-128.ibm.com/developerworks/power/cell/downloads.html [2] http://www.opensource.org/licenses/cpl1.0.php ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3/Cell BE platform
On 11/15/06, Daniel O'Connor <[EMAIL PROTECTED]> wrote: On Thursday 16 November 2006 16:22, Robert McGwier wrote: > The Nividia GPU's have fft and blas running on them. They are doing > teraflops and the tools/SDK are available under NDA. They do indeed > have multiply , accumulate, etc. ATI/AMD have something similar.. http://ati.amd.com/companyinfo/researcher/resources.html (Not that I have used either) And somebody even did the implementation of the fft on a GPU for gnuradio already: www.ecsl.cs.sunysb.edu/fir/fir.ps Looks kind of old, but very interesting... Thomas ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3/Cell BE platform
On Thursday 16 November 2006 16:22, Robert McGwier wrote: > The Nividia GPU's have fft and blas running on them. They are doing > teraflops and the tools/SDK are available under NDA. They do indeed > have multiply , accumulate, etc. ATI/AMD have something similar.. http://ati.amd.com/companyinfo/researcher/resources.html (Not that I have used either) -- Daniel O'Connor software and network engineer for Genesis Software - http://www.gsoft.com.au "The nice thing about standards is that there are so many of them to choose from." -- Andrew Tanenbaum GPG Fingerprint - 5596 B766 97C0 0E94 4347 295E E593 DC20 7B3F CE8C pgpr62deDxpzD.pgp Description: PGP signature ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3/Cell BE platform
The Nividia GPU's have fft and blas running on them. They are doing teraflops and the tools/SDK are available under NDA. They do indeed have multiply , accumulate, etc. We are going to have GnuRadio in good enough shape to "take over the world" when these high speed SDR engines (gaming tools that need DSP hardware) become widely and openly available and as consumer product based, they should be inexpensive. In my personal opinion, the SDR part is now relegated to the "mundane" and the most interesting work is now in AI/Cognitive work, classification, serious mesh networking, etc. that becomes more easily attainable with the high bandwidth channel back to the "front end" which is the completed SDR toolbox operating on these fancy DSP enabled processors with their huge bandwidth IO pipes. The needs of the single-person-shooter games to render video on the fly rather than play through rastered and stored video has certainly been our friend. The "good old days" ARE NOW. Bob Marcus Leech wrote: Jason wrote: All, has anyone been looking at the Cell BE processor as a gnuradio backend platform? For US$600 you get a Playstation 3 that boots linux, and has 7 128bit RISC processors, each with 256k local memory, all moving at ~3.4GHz. Each RISC processor is optimized for vector math / stream processing... :) There is already an SDK (modified GNU toolchain) from IBM. It seems to use a standard toolchain and apps that run on a 64bit ppc, but the toolchain is modified to compile special threads (different exec type) for the RISC processors (called SPEs). threads are written in C against a provided library. My thought for GnuRadio was to make near realtime mod/demod of ATSC a possibility. As well as speeding up most other stream processing. The SDK and info are here [1]. The .iso seems to have everything needed, and the provided code examples aren't too much different from writing pthreads. The magic is in the added instruction set. Good luck getting a hold of one before Christmas :) Jason. Quite apart from the Cell processors, it also boasts dual GPUs, which are also good at doing multiply-accumulate. H. ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio -- AMSAT Director and VP Engineering. Member: ARRL, AMSAT-DL, TAPR, Packrats, NJQRP, QRP ARCI, QCWA, FRC. ARRL SDR WG Chair "You see, wire telegraph is a kind of a very, very long cat. You pull his tail in New York and his head is meowing in Los Angeles. Do you understand this? And radio operates exactly the same way: you send signals here, they receive them there. The only difference is that there is no cat." - Einstein ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3/Cell BE platform
Yes. For a work project my group is purchasing two Mercury 1U Cell servers. This will make ATSC and HDTV a real possibility. These Mercury servers have a high bandwidth PCI express connection for getting data in and out of the world.Another fellow and I will be porting GnuRadio to the Cell BE ont he Mercury. It is not for consumers in this form. These things are over $10,000 a piece with PPC Linux installed. I cannot wait until we get the PS3's hacked to be Linux desktops. Bob Jason wrote: All, has anyone been looking at the Cell BE processor as a gnuradio backend platform? For US$600 you get a Playstation 3 that boots linux, and has 7 128bit RISC processors, each with 256k local memory, all moving at ~3.4GHz. Each RISC processor is optimized for vector math / stream processing... :) There is already an SDK (modified GNU toolchain) from IBM. It seems to use a standard toolchain and apps that run on a 64bit ppc, but the toolchain is modified to compile special threads (different exec type) for the RISC processors (called SPEs). threads are written in C against a provided library. My thought for GnuRadio was to make near realtime mod/demod of ATSC a possibility. As well as speeding up most other stream processing. The SDK and info are here [1]. The .iso seems to have everything needed, and the provided code examples aren't too much different from writing pthreads. The magic is in the added instruction set. Good luck getting a hold of one before Christmas :) Jason. [1] http://www-128.ibm.com/developerworks/power/cell/ ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio -- AMSAT Director and VP Engineering. Member: ARRL, AMSAT-DL, TAPR, Packrats, NJQRP, QRP ARCI, QCWA, FRC. ARRL SDR WG Chair "You see, wire telegraph is a kind of a very, very long cat. You pull his tail in New York and his head is meowing in Los Angeles. Do you understand this? And radio operates exactly the same way: you send signals here, they receive them there. The only difference is that there is no cat." - Einstein ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
Re: [Discuss-gnuradio] PS3/Cell BE platform
Jason wrote: All, has anyone been looking at the Cell BE processor as a gnuradio backend platform? For US$600 you get a Playstation 3 that boots linux, and has 7 128bit RISC processors, each with 256k local memory, all moving at ~3.4GHz. Each RISC processor is optimized for vector math / stream processing... :) There is already an SDK (modified GNU toolchain) from IBM. It seems to use a standard toolchain and apps that run on a 64bit ppc, but the toolchain is modified to compile special threads (different exec type) for the RISC processors (called SPEs). threads are written in C against a provided library. My thought for GnuRadio was to make near realtime mod/demod of ATSC a possibility. As well as speeding up most other stream processing. The SDK and info are here [1]. The .iso seems to have everything needed, and the provided code examples aren't too much different from writing pthreads. The magic is in the added instruction set. Good luck getting a hold of one before Christmas :) Jason. Quite apart from the Cell processors, it also boasts dual GPUs, which are also good at doing multiply-accumulate. H. ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
[Discuss-gnuradio] PS3/Cell BE platform
All, has anyone been looking at the Cell BE processor as a gnuradio backend platform? For US$600 you get a Playstation 3 that boots linux, and has 7 128bit RISC processors, each with 256k local memory, all moving at ~3.4GHz. Each RISC processor is optimized for vector math / stream processing... :) There is already an SDK (modified GNU toolchain) from IBM. It seems to use a standard toolchain and apps that run on a 64bit ppc, but the toolchain is modified to compile special threads (different exec type) for the RISC processors (called SPEs). threads are written in C against a provided library. My thought for GnuRadio was to make near realtime mod/demod of ATSC a possibility. As well as speeding up most other stream processing. The SDK and info are here [1]. The .iso seems to have everything needed, and the provided code examples aren't too much different from writing pthreads. The magic is in the added instruction set. Good luck getting a hold of one before Christmas :) Jason. [1] http://www-128.ibm.com/developerworks/power/cell/ ___ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org http://lists.gnu.org/mailman/listinfo/discuss-gnuradio
