Hello community,
Sorry this update is so late. I spent about a week walking down the
wrong
path so I’ve been spending days and nights trying to get an actual
update.
Logistical:
- Midterm evaluations due tomorrow
- Coproc dev call tomorrow
Progress:
- In getting the FFTC to work, I followed the builtin examples which
require us to run Ti’s code composer software and have the board in a
mode
where the ARMs are shut off. When I tried to program the DSP through the
ARM to run the examples, I get memory errors because the DSP is trying
to
initialize subsystems that have already been initialized by the ARM.
Spent
a long time on that getting nowhere so what Philip recommend to me was
to
try to get things working without the DSP cores starting from hello
world. - I’ve been able to initialize a couple of subsystems like the QMSS and
CPPI. I just need to get the configuration correct for talking to the
FFTC.
I’ve been able to compile the FFTC drivers using the linaro arm
toolchain
so I’m confident this approach will work. TI’s software is multi layered
so
not that many changes are required to port code from the DSP to the ARM. - Taking it a step at a time trying to understand what’s going on is
making
me a better programmer. I feel a lot more prepared to tackle modifying
GNU
Radio. Hopefully I get past setting up the TI chip next week so I can
switch gears.
I’ve been holding back updating the wiki since I wasn’t sure if this
method
was going to get me anywhere but it actually works! will update the wiki
this weekend with a big section on programming this monster using TI’s
libraries and the linaro arm toolchain. I’m sure my method isn’t the
prettiest but hopefully it’s a start so others can develop applications
using the coprocessors through the ARM. I imagine it will be useful to
the
TI people and others using this board.
Not as much progress as I’d hoped but at least I can see the light
again.
Thanks!