Pong

This repository implements Pong in Haskell, both as a software program (using SDL for input and graphics) and as a hardware FPGA design via Clash.

This code is part of the book Retrocomputing with Clash: Haskell for FPGA Hardware Design at https://unsafePerform.IO/retroclash/.

Building

Use stack build to build and stack run pong to run the software version.

For the hardware version, use the included Shakefile. An easy way to run it is via the provided mk shell script. Create a build.mk file with content like the following:

TARGET = nexys-a7-50t
VIVADO_ROOT = /opt/somewhere/where/vivado/is/installed/bin
ISE_ROOT = /opt/somewhere/where/ise/is/installed/ISE/bin/lin64
QUARTUS_ROOT = /opt/somewhere/where/quartus/is/installed/bin

Alternatively, if you have Vivado/ISE installed in Docker or similar, you can create a wrapper script and use that by setting VIVADO instead of VIVADO_ROOT:

TARGET = nexys-a7-50t
VIVADO = /usr/local/lib/docker-scripts/xilinx-vivado-2019.1-ubuntu-18.04/run
ISE = /usr/local/lib/docker-scripts/xilinx-ise-14.7-ubuntu-12.04/run
QUARTUS = /usr/local/lib/docker-scripts/intel-quartus-20.1-ubuntu-20.04/run

Supported target boards

• Xilinx ISE toolchain

  • papilio-one: Papilio One (Spartan-3) with the Arcade MegaBoard
  • papilio-pro: Papilio Pro (Spartan-6) with the Arcade MegaBoard

• Xilinx Vivado toolchain

  • nexys-a7-50t: Nexys A7-50T (Artix-7)

• Intel Quartus toolchain

  • de0-nano: DE0-Nano (Cyclone IV) with Fen Logic VGA666 adapter
  • arrow-deca: Arrow DECA (MAX 10)

• F4PGA (formerly SymbiFlow) open source toolchain

  • nexys-a7-50t.f4pga: Nexys A7-50T (Artix-7)

Bitfile can be built with e.g. ./mk papilio-pro/bitfile. For some targets, a rule to upload the resulting bitfile is also included, e.g. ./mk arrow-deca/upload.

Adding support for other Intel, Xilinx, or F4PGA based FPGA dev boards is very straightforward with the included Shake rules, as long as they have VGA output and at least two input pushbuttons.

Targeting other FPGA toolchains will require adding support in the Shake rules. Alternatively, you can always just run Clash, and import the resulting Verilog files into your FPGA toolchain in a non-automated way.