jab0079 / comp4300-p1 Goto Github PK

Adam Eichelkraut - [email protected]
Jared Brown - [email protected]

This project extended the previous General Purpose Register (GPR) 
multi-cycle machine with a scalar piped architecture into a 
Scoreboard-based controller. Instead of the single pipeline of the
previous simulator, the scoreboard manages four independent functional
units (FUs). The Scoreboard takes care of any potential hazards as well.
For this project, the Scoreboard can be set to either reserve the entire
Functional Unit for each instruction or only the very first stage.
Each instruction has 4 stages: Issue, Fetch Operands, Execution, and Write 
Back. Each specific instruction is encapsulated in its own object that is 
derived from the Instruction class. They take in a Simulator object as 
a context to manipulate. From here the FUs can use these Instructions 
to change properties of the Simulator without having to know the specific 		implementation details.

GENERAL PURPOSE REGISTER MACHINE INSTRUCTION SET (from previous project) 
	
	ADD  -  | 0x01 | 5-bit dest | 5-bit src1 | 5-bit src2 | 9-bit unused |
	ADDI -  | 0x02 | 5-bit dest | 5-bit src1 | 14-bit immediate value... |
	B    -  | 0x03 | 24-bit relative offset ............................ |
	BEQZ -  | 0x04 | 5-bit src1 | 19-bit relative offset ............... |
	BGE  -  | 0x05 | 5-bit src1 | 5-bit src2 | 14-bit offset ........... |
	BNE  -  | 0x06 | 5-bit src1 | 5-bit src2 | 14-bit offset ........... |
	LA   -  | 0x07 | 5-bit src1 | 19-bit relative offset ............... |
	LB   -  | 0x08 | 5-bit src1 | 5-bit src2 | 14-bit offset ........... |
	LI   -  | 0x09 | 5-bit dest | 19-bit immediate value ............... |
	SUBI -  | 0x0A | 5-bit dest | 5-bit src1 | 14-bit immediate value .. |
	NOP  -  | 0x0B | 24-bits usused .................................... |
	SYSCALL | 0x0C | 24-bits usused .................................... |

ADDITIONAL FLOATING POINT REGISTERS (new)

	FADD -  | 0x0D | 5-bit dest | 5-bit src1 | 5-bit src2 | 9-bit unused |
	FMUL -  | 0x0E | 5-bit dest | 5-bit src1 | 5-bit src2 | 9-bit unused |
	FSUB -  | 0x0F | 5-bit dest | 5-bit src1 | 5-bit src2 | 9-bit unused |
	LD   -  | 0x10 | 5-bit dest | 19-bit relative offset ............... |
	SD   -  | 0x11 | 5-bit dest | 19-bit relative offset ............... |

Included with this project are a hierarchy of Makefiles that
should assist with the compilation. To compile, simply type
in 'make' to the commandline while in the uppermost directory
that also contains this README file.

NOTE: The uppermost Makefile creates two directories within
the same directory called /bin and /obj. The /bin directory
will contain all executables for this project. The /obj
directory will contain all compiled objects associated with
their respective source file names.

bin/	  - contains all executables (created by Makefile)

docs/	  - contains all extra documentation and sample code

include/  - contains the header files for all sources

obj/	  - contains all compiled sources (created by Makefile)

src/ 	  - contains all source files

src/stackCode - contains all source files for the stackSim

src/accumCode - contains all source files for the accumSim

src/gprCode - contains all source files for the gprSim

src/pipeCode - contains all source files for the pipeSim

src/scoCode - contains all source files for the scoSim

src/utils     - contains all utility source files 
		            used by all executables

When you run an executable, one of the highlighted
regions to pay attention to is the
"Loading source into memory..." section. This output
happens right before the source gets loaded into memory.

The two blocks of output after that are the USER DATA
and the USER TEXT segments from memory. 
This output is the loaded source file in memory as hexadecimal 
values.

After the USER TEXT segment is when the simulator starts
to run. While the simulator is running, there is no feedback
as to the internal workings from memory during processing.

When the simulator finishes, each sample file records their
output into the USER DATA segment, therefore, we will print
out that segment to show the user the correct answer.

The total clock cycles, total instructions executed, and the
number of NOPs are also displayed at the end of the run and
below.

The program cleans up memory and closes.

Results from lab4c.s (not-pipelined)
Instruction Count: 1331
Cycle Count: 4703
NOP Count: 306

Results from lab4c.s (pipelined)
Instruction Count: 1331
Cycle Count: 1674
NOP Count: 306