How rbsuhail/fpga works
Main Components
• Top Module: Acts as the overall system integrator. It initializes the image RAM, instantiates and interconnects the main modules.
• RAM Module: Stores the 12x12 image (144 bytes). It supports read and write operations for pixel data.
• DataPath Module: Responsible for the core image processing operations, including reading from RAM, performing the Gaussian convolution (using the fixed 5x5 kernel), and writing back processed pixel values. Be sure to include any arithmetic blocks like Booth Multipliers and Division modules which may be used for convolution specifics.
• Additional Verilog Modules: Modules such as bram_1.v, bram_2.v, bram_3.v, booth.v, division.v, ssv_fpga.v and ssv_fpga_top.v should be represented as supporting units that either implement RAM functionalities (or Block RAM interfaces) or provide specialized arithmetic (e.g., multiplication via Booth’s algorithm) required by the DataPath.
Relationships and Interactions
• Control signal routes from the Controller to both the DataPath and the RAM, indicating when read/write operations occur.
• If applicable, show bidirectional connections where the top module provides input signals (like clock, reset) to all subordinate modules.
• Interfaces between the DataPath and arithmetic components (like the Booth multiplier and division units) as sub-components inside the DataPath block.
Architectural Patterns and Design Principles
• Separation of concerns: DataPath handles arithmetic and data processing, while the Controller manages the operation sequencing.
• Pipelining or parallel data processing if relevant (as might be implied by the breakdown of image processing operations in hardware).
Technologies, Frameworks, and Relevant Details
• Mention that the design is implemented in Verilog, targeted for execution on FPGA hardware, using either Xilinx ISE or Vivado.
• Use of Block RAM (BRAM) modules for image storage, possibly detailed by the bram_1, bram_2, and bram_3 files.
• That arithmetic operations (using specialized modules like booth.v for multiplication and division.v for division) are integral for the convolution process with the Gaussian Kernel.