Quiznetik

High Performance Computing (HPC) | Set 4

1. In All-to-All Personalized Communication on a Ring, the size of the message reduces by              at each step

Correct : A. p

2. All-to-All Broadcast and Reduction algorithm on a Ring terminates in                   steps.

Correct : C. p*p

3. In All-to-all Broadcast on a Mesh, operation performs in which sequence?

Correct : B. columnwise, rowwise

4. Messages get smaller in and stay constant in .

Correct : C. scatter, gather

5. The time taken by all-to- all broadcast on a ring is .

Correct : B. t= ts logp + twm(p-1)

6. The time taken by all-to- all broadcast on a mesh is .

Correct : A. t= (ts + twm)(p-1)

7. The time taken by all-to- all broadcast on a hypercube is .

Correct : C. t= 2ts(√p – 1) - twm(p-1)

8. The prefix-sum operation can be implemented using the kernel

Correct : B. one-to-all broadcast

9. Select the parameters on which the parallel runtime of a program depends.

Correct : D. input size

10. The time that elapses from the moment the first processor starts to the moment the last processor finishes execution is called as                       .

Correct : B. overhead runtime

11. Select how the overhead function (To) is calculated.

Correct : C. to = tp - pts

12. What is is the ratio of the time taken to solve a problem on a single processor to the time required to solve the same problem on a parallel computer with p identical processing elements?

Correct : C. scaleup

13. Which is alternative options for latency hiding?

Correct : B. multithreading

14. ______ Communication model is generally seen in tightly coupled system.

Correct : B. shared-address space

15. The principal parameters that determine the communication latency are as follows:

Correct : A. startup time (ts) per-hop time (th) per-word transfer time (tw)

16. The number and size of tasks into which a problem is decomposed determines the __

Correct : A. granularity

17. Average Degree of Concurrency is...

Correct : A. the average number of tasks that can run concurrently over the entire duration of execution of the process.

18. Which task decomposition technique is suitable for the 15-puzzle problem?

Correct : B. exploratory decomposition

19. Which of the following method is used to avoid Interaction Overheads?

Correct : A. maximizing data locality

20. Which of the following is not parallel algorithm model

Correct : D. the speculative model

21. Nvidia GPU based on following architecture

Correct : B. simd

22. What is Critical Path?

Correct : A. the length of the longest path in a task dependency graph is called the critical path length.

23. Which decompositioin technique uses divide-andconquer strategy?

Correct : A. recursive decomposition

24. Consider Hypercube topology with 8 nodes then how many message passing cycles will require in all to all broadcast operation?

Correct : D. the number of maximum nodes level in graph.

25. Scatter is ____________.

Correct : C. one to all personalised communication

26. If there is 4X4 Mesh Topology ______ message passing cycles will require complete all to all reduction.

Correct : C. 8

27. Following issue(s) is/are the true about sorting techniques with parallel computing.

Correct : B. where to store output sequence is the issue

28. Partitioning on series done after ______________

Correct : C. global arrangement

29. In Parallel DFS processes has following roles.(Select multiple choices if applicable)

Correct : A. donor

30. Suppose there are 16 elements in a series then how many phases will be required to sort the series using parallel odd-even bubble sort?

Correct : D. 15

31. Which are different sources of Overheads in Parallel Programs?

Correct : C. all mentioned options

32. The ratio of the time taken to solve a problem on a parallel processors to the time required to solve the same problem on a single processor with p identical processing elements.

Correct : B. the ratio of the time taken to solve a problem on a single processor to the time required to solve the same problem on a parallel computer with p identical processing elements

33. CUDA helps do execute code in parallel mode using __________

Correct : B. gpu

34. In thread-function execution scenario thread is a ___________

Correct : B. worker

35. In GPU Following statements are true

Correct : C. all the mentioned options.

36. Computer system of a parallel computer is capable of_____________

Correct : A. decentralized computing

37. In which application system Distributed systems can run well?

Correct : A. hpc

38. A pipeline is like .................... ?

Correct : A. an automobile assembly line

39. Pipeline implements ?

Correct : D. all of above

40. A processor performing fetch or decoding of different instruction during the execution of another instruction is called ______ ?

Correct : B. pipe-lining

41. VLIW stands for ?

Correct : A. very long instruction word

42. Which one is not a limitation of a distributed memory parallel system?

Correct : B. cache coherency

43. Which of these steps can create conflict among the processors?

Correct : B. concurrent write

44. Which one is not a characteristic of NUMA multiprocessors?

Correct : D. processors access their independent local memories

45. Which of these is not a source of overhead in parallel computing?

Correct : B. less local memory requirement in distributed computing

46. Systems that do not have parallel processing capabilities are?

Correct : A. sisd

47. How does the number of transistors per chip increase according to Moore ´s law?

Correct : D. exponentially

48. Parallel processing may occur?

Correct : C. both[a] and [b]

49. To which class of systems does the von Neumann computer belong?

Correct : D. sisd (single instruction single data)

50. Fine-grain threading is considered as a ______ threading?

Correct : A. instruction-level

51. Multiprocessor is systems with multiple CPUs, which are capable of independently executing different tasks in parallel. In this category every processor and memory module has similar access time?

Correct : A. uma

52. For inter processor communication the miss arises are called?

Correct : B. coherence misses

53. NUMA architecture uses _______in design?

Correct : D. distributed memory

54. A multiprocessor machine which is capable of executing multiple instructions on multiple data sets?

Correct : C. mimd

55. In message passing, send and receive message between?

Correct : A. task or processes

56. The First step in developing a parallel algorithm is_________?

Correct : A. to decompose the problem into tasks that can be executed concurrently

57. The number of tasks into which a problem is decomposed determines its?

Correct : A. granularity

58. The length of the longest path in a task dependency graph is called?

Correct : A. the critical path length

59. The graph of tasks (nodes) and their interactions/data exchange (edges)?

Correct : A. is referred to as a task interaction graph

60. Mappings are determined by?

Correct : C. both a and b

61. Decomposition Techniques are?

Correct : D. all of above

62. The Owner Computes Rule generally states that the process assigned a particular data item is responsible for?

Correct : A. all computation associated with it

63. A simple application of exploratory decomposition is_?

Correct : A. the solution to a 15 puzzle

64. Speculative Decomposition consist of _?

Correct : C. both a and b

65. task characteristics include?

Correct : D. all of above

66. Writing parallel programs is referred to as?

Correct : D. parallel programming

67. Parallel Algorithm Models?

Correct : A. data parallel model

68. The number and size of tasks into which a problem is decomposed determines the?

Correct : A. fine-granularity

69. A feature of a task-dependency graph that determines the average degree of concurrency for a given granularity is its ___________ path?

Correct : A. critical

70. The pattern of___________ among tasks is captured by what is known as a task-interaction graph?

Correct : A. interaction

71. Interaction overheads can be minimized by____?

Correct : A. maximize data locality

72. Type of parallelism that is naturally expressed by independent tasks in a task-dependency graph is called _______ parallelism?

Correct : A. task

73. Speed up is defined as a ratio of?

Correct : A. s=ts/tp

74. Parallel computing means to divide the job into several __________?

Correct : D. task

75. _________ is a method for inducing concurrency in problems that can be solved using the divide-and-conquer strategy?

Correct : C. data-decomposition

76. The___ time collectively spent by all the processing elements Tall = p TP?

Correct : A. total

77. The dual of one-to-all broadcast is ?

Correct : A. all-to-one reduction

78. A hypercube has?

Correct : A. 2d nodes

79. The Prefix Sum Operation can be implemented using the ?

Correct : A. all-to-all broadcast kernel.

80. In the scatter operation ?

Correct : A. single node send a unique message of size m to every other node

81. The gather operation is exactly the inverse of the ?

Correct : A. scatter operation

82. Parallel algorithms often require a single process to send identical data to all other processes or to a subset of them. This operation is known as _________?

Correct : A. one-to-all broadcast

83. In which of the following operation, a single node sends a unique message of size m to every other node?

Correct : D. both a and c

84. Gather operation is also known as ________?

Correct : A. one to all personalized communication

85. Conventional architectures coarsely comprise of a?

Correct : D. all of above

86. Data intensive applications utilize?

Correct : A. high aggregate throughput

87. A pipeline is like?

Correct : A. overlaps various stages of instruction execution to achieve performance.

88. Scheduling of instructions is determined?

Correct : D. all of above

89. VLIW processors rely on?

Correct : A. compile time analysis

90. Memory system performance is largely captured by?

Correct : C. both a and b

91. The fraction of data references satisfied by the cache is called?

Correct : A. cache hit ratio

92. A single control unit that dispatches the same Instruction to various processors is?

Correct : A. simd

93. The primary forms of data exchange between parallel tasks are?

Correct : C. both a and b

94. The First step in developing a parallel algorithm is?

Correct : A. to decompose the problem into tasks that can be executed concurrently

95. The Owner Computes Rule generally states that the process assigned a particular data item are responsible for?

Correct : A. all computation associated with it

96. A simple application of exploratory decomposition is?

Correct : A. the solution to a 15 puzzle

97. Speculative Decomposition consist of ?

Correct : C. both a and b

98. Task characteristics include?

Correct : D. all of above.

99. The dual of one-to-all broadcast is?

Correct : A. all-to-one reduction

100. A hypercube has?

Correct : A. 2d nodes