Computer-Architecture

Created a CPU in Verilog for ECE251 Final Project.

Rob Website

https://github.com/robmarano/cpu_hdl_catalog/tree/STEP-01

OttoBit Spreadsheet

https://cooperunion.sharepoint.com/:x:/r/sites/Section_ECE-251-A-2023SP/_layouts/15/Doc.aspx?action=edit&sourcedoc=%7B2cd312b1-8aa7-483c-a74d-31bccdc8a4b7%7D&wdOrigin=TEAMS-WEB.teamsSdk.openFilePreview&wdExp=TEAMS-CONTROL&web=1

ISA for Besties CPU

^{Datapath for our CPU.}

^{Register, Immediate, and Address Instruction Types Formating}

^{The instructions we defined for R-type and how their bit values are defined.}

^{The instructions we defined for I-type and how their bit values are defined.}

^{The instructions we defined for J-type and how their bit values are defined.}

1. The ALU Operand Size is 32 bits.
2. We'd support 12 ALU operations: add, subtract, shift left, shift right, and, or, xor, nor, nand, nor, jr, set less than, and set greater than.
3. The Operational Code (OpCode) will be 5 bits.
4. We'll provide 128 registers for the programs (7 bits for register addressing).
5. We will have a return address to receive the results of the ALU ($ra, reg #127)
6. We'll have 3 types of instruction format: Register (R), Immediate (I), & Jump (J)
7. Our instruction width will be a fixed value of 32 bits.
8. We'll design our datapath and Control Unit similar to MIPs but make any changes we want/need to make.
9. The datapath and control unit will be implemented using a single-cycle instruction execution since will plan on using RISC.
10. The single-cycle implementation won't be converted to a pipelined design.
11. Our Verilog Components include: Adder, Alu, AluDecoder, Clock, Computer, Controller, Cpu, D-Flip Flops, Datapath, Dmem, FullAdder, Imem, MainDecoder, Mux (2:1), RegisterFile, SignExtender, and Sl2 (Shift left 2)

Other Decisions Made

• First two bits for OpCode determine instruction type (01 or 10 for Register, 11 for Immediate, 00 for Jump)
• We'll be storing our numbers using Little Endianness
• Registers are written on rising edge
• For our code, n=32 bits and r=7 bits. n is for the operand bit size and r represents the number of bits we have for our registers.