DigitalLogic ComputerOrganization L26 Revision
DigitalLogic ComputerOrganization L26 Revision
COMPUTER ORGANIZATION
Lecture 26: REVISION
ELEC3010
FINAL EXAM
❑Closed book
▪ You can use 1 A4 page cheat sheet and 1 A4 Page listing all assembly
instructions of the processor discussed in class
❑Topics covered:
▪ Convert from C to Assembly
▪ Single cycle/Pipelined/Multi-core processors
▪ Caches
▪ Virtual memory
▪ IO
▪ Verilog
❑ You have 100 minutes to earn 100 points
2
CACHES
3
CACHES
❑Find the AMAT given that the L1 cache access time is 1 ns,
L2 cache access time is 10 ns, memory access time is 100 ns,
L1 cache miss rate is 5%, L2 cache local miss rate is 10%.
a. 1 ns
b. 2 ns
c. 3 ns
d. 4 ns
5
CACHES
6
MULTI-CORE
Given a program with a serial part and a parallel part. Assume that it
takes 20ms to execute the serial part and 20ms to execute the parallel
part on a 1-core processor. What will be the execution time of the
program and the corresponding speedup if the number of cores is 4?
A. 10ms and 4
B. 20ms and 2
C. 25ms and 1.6
D. 25ms and 1.8
7
MULTI-CORE
A. 56 seconds
B. 70 seconds
C. 35 seconds
D. 60 seconds
8
MULTI-CORE
a) How many threads can the processor execute at the same time?
b) How many instructions can the processor execute at the same time?
9
VIRTUAL MEMORY
A. 16 MB
B. 4 MB
C. 32 MB
D. 8 MB
11
VIRTUAL MEMORY
The following problems concern the way virtual addresses are translated into physical addresses.
12
ELEC3010 IN ONE SLIDE
C int x = 10; Interrupts system
x = 2* x + 15; I/O reserved PAGE
compiler stack TABLE
sub R5, R5, R5
addi R5, R5, 10
Assembly
shl R5, R5, 2
assembler addi R5, R5, 15 heap
00000000101000000000001010010100 global
machine 00000000101000000000001010010011 text
code 00000000001000101000001010000000
00000000111100101000001010010011
loader system reserved
CPU FSMs
Circuits RF
A Architecture -- ISA
L 32 32 Microarchitecture
Gates A U
B • Pipelining
• Caches
Transistors
HDL Parallelism:
• ILP KERNEL
Silicon
• TLP/Multicore
13