I know about Jacobi algorithm and Givens rotation but I don't know how to implement it on FPGA and write a code in VHDL ( or even Verilog).
Can anyone help?
If you don't need the blazing speed, the fastest way to get it up and running might be to implement a soft-processor and just run C/C++ code doing all the math on it.
It depends of many factors such as matrix size, numbers range and precision, speed, target FPGA, etc.
There are several companies who manufacture DSP / FPGA cards with which to do this but I must warn they are quite expensive. For example Xilinx have an FPGA blockset which integrates with MATLAB / SIMULINK and which allows one to both simulate and run any FPGA designs in real-time. I would suggest the following to start
1 Xilinx - development boards and IDE suite.
2 Innovative Integration - www.innovative.dsp.com
3 dSPACE - http://www.dspace.com/en/inc/home/products/hw/modular_hardware_introduction/i_o_boards/ds5203_fpga_board.cfm
I have attached an II product brochure. But as I say there are many others. Some FPGA manufacturers, like II, allow C, C++ ---> FPGA co-development development as well.
Sorry but it would not allow me to upload another brochure for dSPACE. It appears ResearchGate place a limit on what may be uploaded per topic.
These tools will free you from having to get to learn basic FPGA programming as they are higher level FPGA development tools to allow you to focus on implementing the algorithms, which are quite procedural in themselves.
The most efficient way would be to program your algorithm using C/C++ and generate HDL code (VHDL/Verilog) using an HLS tool. The less expensive one is from Xilinx: VIVADO HLS. If you work for an university you can get the tools for free throughthe XUP (Xilinx University Program).
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Optics