Running Python on an FPGA

[AyaZab]

Hello, there are 3 FPGAs I am interested in, but I'd like to clear up a technical doubt first.

- Intel DE10 Nano FPGA

- Intel Cyclone 10GX FPGA

- Intel Arria 10GX FPGA

Is it possible to run Python code on any of these devices? If so, how can that be done?

Quick responses would be greatly appreciated, thank you very much.

[_AK6DN_]

First, "DE10 Nano" is not an FPGA per se, it is a development board with a CycloneV 5CSEBA6 device with a dual core ARM HPS.
TerASIC (the maker) has ported linux to it, so running Python on the ARM cpu is a no brainer.
Refer to: https://www.terasic.com.tw/cgi-bin/page/archive.pl?CategoryNo=165&No=1046&PartNo=1

Cyclone10GX is more complicated, there is no embedded CPU option. So you would have to use a soft core CPU design.

Arria10GX is similar to Cyclone10GX; no embedded hard CPU option. So again a soft core CPU design would be required.

OTOH the Arria10SX variation has an embedded dual core ARM hard CPU.
TerASIC has a development board with a linux port for this board also.
Refer to: https://www.terasic.com.tw/cgi-bin/page/archive.pl?CategoryNo=216&No=997&PartNo=1

Do you expect to run Python code on a basic FPGA without either an embedded hard CPU or a soft core CPU?

There is this:  https://pycpu.wordpress.com/ which basically is small CPU design that runs a subset of Python code
in an FPGA. It translates Python to an intermediate form and uses a soft core processor to run the Python code.

Lastly, if what you want to run is Python programs, why use an FPGA at all? A small RaspberryPi or equivalent CPU board
would do just as well and be a lot more cost effective and easier to use. Python programming and FPGA programming
are really two very different subjects and not really related.

[ShengN_Intel]

Hi,

Do you have any further update or concern?

Thanks,

Best Regards,

Sheng

[Sidhharth1310]

Hi Sheng,

I have a similar question. My FPGA I have is a DE1 SoC cyclone V  and I'm also planning to use python to run my pre trained ML model using the ARM Cortex A9 processor. The reason is I'm already using the FPGA fabric to create a custom FFT core where the FFT core's results are used as input in the python code (ML model maybe). I want to know if this is possible with my setup and how to achieve this.

Kind regards,

Sidhharth