Test your basic knowledge |

Computer Architecture And Design

Instructions:

Answer 38 questions in 15 minutes.
If you are not ready to take this test, you can study here.
Match each statement with the correct term.
Don't refresh. All questions and answers are randomly picked and ordered every time you load a test.

This is a study tool. The 3 wrong answers for each question are randomly chosen from answers to other questions. So, you might find at times the answers obvious, but you will see it re-enforces your understanding as you take the test each time.

1. Amdahl's Law
The most expensive computers - costing tens of millions of dollars. They emphasize floating- point performance.
Procedure call. Copies PC to $ra - # push $t0 - Subu $sp - $sp - 4 - sw $t0 - ($sp) - # pop $t0 - Lw $t0 - ($sp) - addu $sp - $sp - 4
The performance enhancement possible with a given improvement is limited by the amount that the improvement feature is used.
1- response time 2- throughput - response time and throughput are directly proportional or only interrelated - Interrelated only.

2. What is response time?
The total time required for the computer to to complete a task. (Includes disk accesses - memory accesses - I/O activities - OS overhead - and CPU execution time.)
When a segment of the application has an absolute maximum execution time.
Desktop computer / laptop computer - server - super computer - embedded computer - mobile computing
An abstract interface between the hardware and the lowest level software that encompasses all the information necessary to write a machine language program that will run correctly - including instructions - registers - memory access - I/O - etc.

3. Moore's Law
Also called ILP. This is the potential overlap among instructions. There are two approaches: (1) hardware - and (2) software.
Computer speeds double every 18-24 months
Computers that are lodged in other devices where their presence is not immediately obvious.
The most expensive computers - costing tens of millions of dollars. They emphasize floating- point performance.

4. What are the industry standard benchmarks to measure performance (e.g. - with different vendor chips)?
1- response time 2- throughput - response time and throughput are directly proportional or only interrelated - Interrelated only.
The total time required for the computer to to complete a task. (Includes disk accesses - memory accesses - I/O activities - OS overhead - and CPU execution time.)
The combination of performance (measured primarily in therms of compute performance and graphics performance) and the price of a system.
(1) pipelining (2) multiple instruction issue

5. What is the $epc register used for?
A faster processor to complete task sooner - a better algorithm to complete the program/task sooner.
Desktop computer / laptop computer - server - super computer - embedded computer - mobile computing
Points to the address of an instruction that caused an exception
Magnetic disk - flash memory are examples of this type of memory.

6. What are the base units of GHz?
Instructions/unit time (e.g. - instructions/sec) - equal to 1/execution time
Dedicated argument registers to reduce stack usage during procedure calls - consistently sized opcodes - separate instructions for store and load - improved linkage (jal and jr save $ra without using stack)
10^9 cycles per sec
1- response time 2- throughput - response time and throughput are directly proportional or only interrelated - Interrelated only.

7. What are two examples of instruction - level parallelism?
Procedure call. Copies PC to $ra - # push $t0 - Subu $sp - $sp - 4 - sw $t0 - ($sp) - # pop $t0 - Lw $t0 - ($sp) - addu $sp - $sp - 4
When it is possible to occasionally miss the time constraint on an event - as long as not too many are missed.
Using fixed or variable length encoding.
(1) pipelining (2) multiple instruction issue

8. What is price performance?
The combination of performance (measured primarily in therms of compute performance and graphics performance) and the price of a system.
Instructions/unit time (e.g. - instructions/sec) - equal to 1/execution time
Also called DLP. A form of parallelization of computing across multiple processors in parallel computing environments - which focuses on distributing the data across different parallel computing nodes.
The performance enhancement possible with a given improvement is limited by the amount that the improvement feature is used.

9. An example of an improvement that would impact throughput (but not response time).
Add memory - additional processors to handle more tasks in a given time.
Dedicated argument registers to reduce stack usage during procedure calls - consistently sized opcodes - separate instructions for store and load - improved linkage (jal and jr save $ra without using stack)
Points to the current top of the stack
Storage that retains data even in the absence of a power source.

10. How is CPU performance measured?
Instructions/unit time (e.g. - instructions/sec) - equal to 1/execution time
Memory used to hold program while they are executing.
Points to the address of an instruction that caused an exception
Also called TLP. A form of parallelization of computer code across multiple processors in parallel computing environments - which focuses on distributing execution processes (threads) across different parallel computing nodes.

11. What are the five classic components of a computer?
Storage that retains data even in the absence of a power source.
There does not exist the case of negative zero. - Can perform a- b as a+ (- b) without adjustments inside the CPU.
Non - volatile memory used to store programs and data between executions.
Input - output - memory - datapath - control

12. What is soft real- time?
When it is possible to occasionally miss the time constraint on an event - as long as not too many are missed.
Memory used to hold program while they are executing.
The combination of performance (measured primarily in therms of compute performance and graphics performance) and the price of a system.
Input - output - memory - datapath - control

13. What is an Instruction Set Architecture (ISA)?
An abstract interface between the hardware and the lowest level software that encompasses all the information necessary to write a machine language program that will run correctly - including instructions - registers - memory access - I/O - etc.
Also called DLP. A form of parallelization of computing across multiple processors in parallel computing environments - which focuses on distributing the data across different parallel computing nodes.
Using fixed or variable length encoding.
Points to the address of an instruction that caused an exception

14. How can you encode an ISA?
(1) pipelining (2) multiple instruction issue
Using fixed or variable length encoding.
Input - output - memory - datapath - control
Computers that are lodged in other devices where their presence is not immediately obvious.

15. What is the $pc register used for?
The combination of performance (measured primarily in therms of compute performance and graphics performance) and the price of a system.
Using fixed or variable length encoding.
Points to the next instruction to be executed.
Non - volatile memory used to store programs and data between executions.

16. What is thread- level parallelism?
Computers that are lodged in other devices where their presence is not immediately obvious.
DRAM - RAM - Cache are examples of this type of memory.
Also called TLP. A form of parallelization of computer code across multiple processors in parallel computing environments - which focuses on distributing execution processes (threads) across different parallel computing nodes.
Magnetic disk - flash memory are examples of this type of memory.

17. What does hardware refer to?
Storage that retains data only if it is receiving power
The specifics of a computer - including the detailed logic design and the packaging technology of the computer
The combination of performance (measured primarily in therms of compute performance and graphics performance) and the price of a system.
1- response time 2- throughput - response time and throughput are directly proportional or only interrelated - Interrelated only.

18. What is instruction - level parallelism?
Storage that retains data even in the absence of a power source.
Add memory - additional processors to handle more tasks in a given time.
Points to the current top of the stack
Also called ILP. This is the potential overlap among instructions. There are two approaches: (1) hardware - and (2) software.

19. What is non - volatile memory?
Desktop computer / laptop computer - server - super computer - embedded computer - mobile computing
Also called TLP. A form of parallelization of computer code across multiple processors in parallel computing environments - which focuses on distributing execution processes (threads) across different parallel computing nodes.
1- response time 2- throughput - response time and throughput are directly proportional or only interrelated - Interrelated only.
Storage that retains data even in the absence of a power source.

20. What are embedded computers?
When it is possible to occasionally miss the time constraint on an event - as long as not too many are missed.
Computers that are lodged in other devices where their presence is not immediately obvious.
There does not exist the case of negative zero. - Can perform a- b as a+ (- b) without adjustments inside the CPU.
Desktop computer / laptop computer - server - super computer - embedded computer - mobile computing

21. An example of volatile memory
The specifics of a computer - including the detailed logic design and the packaging technology of the computer
Instructions and data are stored in memory as numbers
Computer speeds double every 18-24 months
DRAM - RAM - Cache are examples of this type of memory.

22. What is a supercomputer?
The most expensive computers - costing tens of millions of dollars. They emphasize floating- point performance.
When a segment of the application has an absolute maximum execution time.
Desktop computer / laptop computer - server - super computer - embedded computer - mobile computing
There does not exist the case of negative zero. - Can perform a- b as a+ (- b) without adjustments inside the CPU.

23. What does jal <proc> do?
Desktop computer / laptop computer - server - super computer - embedded computer - mobile computing
Points to the address of an instruction that caused an exception
The combination of performance (measured primarily in therms of compute performance and graphics performance) and the price of a system.
Procedure call. Copies PC to $ra - # push $t0 - Subu $sp - $sp - 4 - sw $t0 - ($sp) - # pop $t0 - Lw $t0 - ($sp) - addu $sp - $sp - 4

24. What is secondary memory?
Using fixed or variable length encoding.
Non - volatile memory used to store programs and data between executions.
Procedure call. Copies PC to $ra - # push $t0 - Subu $sp - $sp - 4 - sw $t0 - ($sp) - # pop $t0 - Lw $t0 - ($sp) - addu $sp - $sp - 4
There does not exist the case of negative zero. - Can perform a- b as a+ (- b) without adjustments inside the CPU.

25. What is main/primary memory?
An abstract interface between the hardware and the lowest level software that encompasses all the information necessary to write a machine language program that will run correctly - including instructions - registers - memory access - I/O - etc.
Memory used to hold program while they are executing.
The most expensive computers - costing tens of millions of dollars. They emphasize floating- point performance.
Instructions/unit time (e.g. - instructions/sec) - equal to 1/execution time

26. What are the classes of computing applications (five)?
10^9 cycles per sec
Desktop computer / laptop computer - server - super computer - embedded computer - mobile computing
Also called DLP. A form of parallelization of computing across multiple processors in parallel computing environments - which focuses on distributing the data across different parallel computing nodes.
When a segment of the application has an absolute maximum execution time.

27. An example of non - volatile memory
Also called ILP. This is the potential overlap among instructions. There are two approaches: (1) hardware - and (2) software.
There does not exist the case of negative zero. - Can perform a- b as a+ (- b) without adjustments inside the CPU.
Storage that retains data only if it is receiving power
Magnetic disk - flash memory are examples of this type of memory.

28. An example of an improvement that would impact response time (but not throughput).
Algorithm - programming language - compiler - instruction set architecture
A faster processor to complete task sooner - a better algorithm to complete the program/task sooner.
Dedicated argument registers to reduce stack usage during procedure calls - consistently sized opcodes - separate instructions for store and load - improved linkage (jal and jr save $ra without using stack)
The number of tasks completed per unit of time.

29. What is included in the term organization?

30. What is volatile memory?
Points to the current top of the stack
10^9 cycles per sec
Dedicated argument registers to reduce stack usage during procedure calls - consistently sized opcodes - separate instructions for store and load - improved linkage (jal and jr save $ra without using stack)
Storage that retains data only if it is receiving power

31. What is data- level parallelism?
Add memory - additional processors to handle more tasks in a given time.
The performance enhancement possible with a given improvement is limited by the amount that the improvement feature is used.
When it is possible to occasionally miss the time constraint on an event - as long as not too many are missed.
Also called DLP. A form of parallelization of computing across multiple processors in parallel computing environments - which focuses on distributing the data across different parallel computing nodes.

32. What is throughput?
There does not exist the case of negative zero. - Can perform a- b as a+ (- b) without adjustments inside the CPU.
Instructions and data are stored in memory as numbers
Desktop computer / laptop computer - server - super computer - embedded computer - mobile computing
The number of tasks completed per unit of time.

33. One reason why two's compliment is used as opposed to signed magnitude or one's compliment?
There does not exist the case of negative zero. - Can perform a- b as a+ (- b) without adjustments inside the CPU.
Using fixed or variable length encoding.
Procedure call. Copies PC to $ra - # push $t0 - Subu $sp - $sp - 4 - sw $t0 - ($sp) - # pop $t0 - Lw $t0 - ($sp) - addu $sp - $sp - 4
DRAM - RAM - Cache are examples of this type of memory.

34. What is the $sp register used for?
Add memory - additional processors to handle more tasks in a given time.
Magnetic disk - flash memory are examples of this type of memory.
Points to the current top of the stack
Using fixed or variable length encoding.

35. What is a real- time performance requirement?
There does not exist the case of negative zero. - Can perform a- b as a+ (- b) without adjustments inside the CPU.
Also called DLP. A form of parallelization of computing across multiple processors in parallel computing environments - which focuses on distributing the data across different parallel computing nodes.
When a segment of the application has an absolute maximum execution time.
Instructions/unit time (e.g. - instructions/sec) - equal to 1/execution time

36. What are the hardware/software components affecting program performance?
(1) pipelining (2) multiple instruction issue
Algorithm - programming language - compiler - instruction set architecture
Magnetic disk - flash memory are examples of this type of memory.
The combination of performance (measured primarily in therms of compute performance and graphics performance) and the price of a system.

37. An example of something typically associated with RISC architecture that is not typical in CISC architecture.
Dedicated argument registers to reduce stack usage during procedure calls - consistently sized opcodes - separate instructions for store and load - improved linkage (jal and jr save $ra without using stack)
(1) pipelining (2) multiple instruction issue
Storage that retains data even in the absence of a power source.
Memory used to hold program while they are executing.

38. Stored Program Concept
Instructions and data are stored in memory as numbers
Points to the address of an instruction that caused an exception
The specifics of a computer - including the detailed logic design and the packaging technology of the computer
Also called TLP. A form of parallelization of computer code across multiple processors in parallel computing environments - which focuses on distributing execution processes (threads) across different parallel computing nodes.