Vincenzo P. wrote:
really thanks in advance
Have you been the gnuradio trac pages on installing FC7 and GnuRadio on
I run the FM Stereo as a demonstration remotely (so the desktop code is
not taking extremely scarce main memory) and it runs just fine. This is
one of the more complex pieces of code. The new OFDM code (still in
branch and not checked in) would be another.
NONE OF THESE use the SPE’s yet but they run on the PPU’s (the Power
We are definitely working on the PS3 in a big way.
I am happy to announce that today I completed the UPGRADE from FC7 to
Fedora 8. I first uninstalled CellSDK 3.0, did the upgrade, and then
installed. What I was checking was to see if the install would proceed
without complaint. It proceeded to completion and I have built gnuradio
using CellSDK 3.0 running under Fedora 8.
The only trouble I ran into at all was caused by not carefully checking
the FC7 installation page before I followed it blindly. If you followed
those instructions some time ago and added the fedora development
repository, you will find that your system is littered with FC8 and FC9
candidate packages. This took me a day to clean off but then the
upgrade proceeded without incident after I managed to return the state
of the code to full F7 compliant. (ARRRRGGGHHH).
I ran into a problem attempting to BOOT the F8 PPC DVD iso. It fails
utterly. Since I could not see a way around this, I proceeded to do the
update. That would have taken a couple of hours, primarily downloading
rpm’s, had it not been for the mistake in recommending the development
F8 runs just fine with a customized PS3 kernel. I have customized it to
allow NFS mounting for cross compile and to remove LOTS of memory
hogging junk that will never be on the PS3.
The kernel I run on the dual blade machines at work is necessarily
different because it has the Axon 30 or 40 Northbridge, all spe’s and
the ability to support cards through firewire, USB, PCIe, serial ports,
In all cases, the only way to go on the PS3 is cross compile and to
install a minimal kernel after it is up and running. This saves memory
on the PS3 (only 256 MB) and reduces the compile DRAMATICALLY.
On tests run on the PS3, it is extremely capable of handling very
wideband signals IF WE CAN GET THEM INTO THE MACHINE. We will be able
to do many Mbaud signals, well into the tens of Mbaud. HDTV should be
a cakewalk and barely exercise the processor once that code is written.
There is a problem currently that I can only assume will be remedied
over time. The Gaelic GigE driver STINKS. It can barely sustain 100
baseT full duplex rates. I understand there is a proprietary driver
that shows great speed but I cannot predict how long this will take for
the existence of this to have an impact on the open source driver. The
complaints about its existence are real enough!
The FFTW support for the Cell is weak. You can see this in the
benchmarks of the FFT on the FFTW Cell alpha page. The numbers are way
off base. The cell is capable of much more, even with six SPE’s only.
Eric is looking carefully at this with some initial indication that much
more can be done.
This is getting attention and it is my belief we will be delivering some
very nice apps in the next few months for the PS3. I make no
predictions on which or when.
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