This basic CPU runs on Spartan 3E FPGA with basic instructions. It is Von Neumann architecture. Each instruction is 32 bit. ALU(Arithmetic Logic Unit) is capable of 32-bit if operands are loaded from memory. Otherwise direct operations only support 14-bit. My github repo has all necessary files (Design files, binary download, memory dump and example bubble sort assembly file). Unfortunately no interrupt vector support yet. But I will add that soon.

Instruction Word Bit Field Breakdown

| opcode  |   i   |     A     |     B     |
|  3-bit  | 1-bit |   14-bit  |   14-bit  |

ADD -> unsigned add

{opcode, i} = {0, 0}
*A <- (*A) + (*B)

ADDi -> unsigned add immediate

{opcode, i} = {0, 1}
*A <- (*A) + B

MUL -> unsigned multiply

{opcode, i} = {7, 0}
*A <- (*A) * (*B)

MULi -> unsigned multiply

{opcode, i} = {7, 1}
*A <- (*A) * B

NAND -> bitwise NAND

{opcode, i} = {1, 0}
*A <- ~((*A) & (*B))

NANDi -> bitwise NAND immediate

{opcode, i} = {1, 1}
*A <- ~((*A) & B)

SRL -> shift right if the shift amount *B is less than 32, otherwise shift left

{opcode, i} = {2, 0}
*A <- (*B) < 32 ? (*A) >> (*B) : (*A) << ((*B) – 32)

SRLi -> Shift Right if the shift amount B is less than 32, otherwise Shift Left

{opcode, i} = {2, 1}
*A <- B < 32 ? (*A) >> B : (*A) << (B – 32)

LT -> if *A is less than *B then *A is set to 1, otherwise to 0.

{opcode, i} = {3, 0}
*A <- (*A) < (*B)

LTi -> if *A is less than B then *A is set to 1, otherwise to 0.

{opcode, i} = {3, 1}
*A <- (*A) < B

CP -> copy *B to *A

{opcode, i} = {4, 0}
*A <- *B

CPi -> copy B to *A

{opcode, i} = {4, 1}
*A <- B

CPI -> copy **B to *A

{opcode, i} = {5, 0}
*A <- **B

CPIi -> copy *B to **A

{opcode, i} = {5, 1}
**A <- *B

BZJ -> branch on zero

{opcode, i} = {6, 0}
PC <- (*B) == 0 ? (*A) : (PC+1)

BZJi -> jump

{opcode, i} = {6, 1}
PC <- (*A) + B

Bubble sort algorithm written in this assembly code: bubble_sort.asm

References

• The instructions are outlined by H. Fatih Ugurdag in one his graduate FPGA design courses.