compling programs for the x86 processor for the Parallella

[Lou_Reed]

I have come across many programs that I would lik e to run on the Parallella. They are in the downloadable form *.tgz.

From that one must unpack and the go to to the main directory and then type the sequence

./configure ./make and ./make install.

These programs are clearly for the x86 systems. But since they are
packaged as source can one use them on the Parallella, by simply recompiling
and using the three sequence command as describer above?

It would be nice if we could do that.

Any help appreciated?

Thanks in advance.

Respectfully,

Lou_Reed

[9600]

[rowan194]

Note that when you compile a generic program, it will only execute on the ARM CPU (not the 16/64 core Epiphany).

Only Epiphany specific source code will execute on that processor.

[sebraa]

Note that using the Ubuntu package system (that is, using apt-get, aptitude, or similar) is usually a much easier way than compiling software from source.

[BOBNINO]

hi

sebraa but i normally get an error message saying that package architecture (i386) does not match system (armhf).

i don't want to force the dpkg just in case it corrupt the system file though , so what's the answer.

[piotr5]

the answer is that you need to install the required dev-packages and then cmpile whatever isn't available for arm -- don't know what commands are needed for that. however, if the package is using cuda or whatever graphics-acceleration features, you'll have the bad luck that you'd need to port it to epiphany first. even mesa/opengl hasn't been ported yet, it's using arm and not epiphany for drawing...

[BOBNINO]

hi

piotr5 thats bad i hope its not the same with OPENCL , IS IT??

[9600]

[BOBNINO]

well Andrew i think the whole point for me at least is to get speed from the epiphny with as little configuration as possibly , i mean if one wanted to programme an FPGA using verilog or HDL or what ever language they are using now days then i could have just bought an FPGA board if you know what i mean.(basically what you saying is that its fast but not easy to use!!!)

it would have been more beneficial for everyone if parallella had a library that fit into a c or c++ environment without porting, OPENCL has a context building functions that look for all devices on a machine (fpga, asics, cpu, gpu) so i hope that it will find the epiphny and use it the classical way (like it finds all the others i mentioned before) once i get it going or its another few hours if not days to figure out the porting.

[piotr5]

well, parallella is an fpga-board! as of yet it is unknown what could be done with the epiphany chip, so it's good that additional possibility exists. but imho your problem stems from the opposite direction: as of yet there exists no true context building lib which automatically alters your sources to fit the underlying architecture -- those programs aren't smart enough as of yet. as I understood fpga is a logic-tree architecture, programs there are composed of parallell commands transforming multiple inputs into a single output. ordinary computers, including parallella, use a von-neumann sequential execution machine, none of the commands really are supposed to be executed in parallell. gpu programming means you write a single sequential program and run the same program on many cores. of course it's possible to port automatically between these 3 systems, but there will always be big differences in the speed and memory situation. you write a program that's optimal on one system and the automatic transformation will produce something that's extremely slow or extremely ressource-hungry on another. you'd need an ai that understands the gist of what you're trying to do and rewrites it for another system. I doubt opencl was designed with that purpose in mind...

anyway, as was said, opencl for parallella isn't ready yet, it isn't in the list of supported hardware for opencl. guess the low-level stuff need to be added in pal before any work on that can start...