Parallel Mids
Parallel Mids
2. In parallel computation, the code fragment provided sums rows of the matrix into a vector
column_sum. (T)
3. In multiplying a matrix with a vector, computing each element of the result as a dot product of
a row of the matrix with the vector is more efficient than multiplying column-by-column. (F)
4. Exploiting spatial and temporal locality in applications has no impact on memory latency or
effective memory bandwidth. (F)
5. The ratio of operations to memory access is not a good indicator of anticipated tolerance to
memory bandwidth. (F)
6. Organizing computation appropriately does not affect spatial and temporal locality. (F)
9. For most parallel algorithms, the per-hop time t ₕ is less than t ʷ *m, even for small values of
m. The total communication cost for a message in a parallel machine is directly proportional to
the number of communication links. (T)
10. In dynamic interconnection networks, switches, and communication links are not used. (F)
11. If the time taken by a processor to access any memory word in the system is identical. the
platform is classified as (UMA). (T)
12. The time is taken to put a memory request and returns a block of data of size P containing the
requested word is referred to as the latency of the memory (T)
13. A parallel Computer is a collection of processors. typically, of the same type. interconnected
in a certain fashion to allow the coordination of their activities and the exchange of data. (T)
14. In the aggressive model instructions can be issued only in the order in which they are
encountered. (F)
15. VLIW processors rely on compile-time analysis to identify and bundle together instructions
that can be executed concurrently (T).
16. In distributed address space platform is ensuring that concurrent operations on multiple
copies of the same memory have well-defined semantics (T).
17. Buses are responsible for transferring data to and from the computer's internal hardware
such as memory. processor. and input and output devices (T).
18. Memory bandwidth can be improved by increasing the size of memory blocks (T)
19. Interactions of processing nodes in the network are accomplished using messages
non-synchronically among the process. during message passing in the network model (F).
20. The memory computer is historically used for architectures in which the memory is physically
shared among various processors. (T)
Q (3) Match the following with their utilization of parallel platforms: (5 Marks]
Application Utilization
1. Automobile processors A. Embedded systems
2. Mail servers B. Large scale servers
3. Network intrusion detection C. Optimizing business and marketing decisions
4. Overlay networks D. Structured peer-to-peer networks
5. Data mining and analysis E. Computer Systems
1. Automobile processors __ A. Embedded systems
2. Mail servers __ B. Large scale servers
3. Network intrusion detection __ E. Computer Systems
4. Overlay networks __ D. Structured peer-to-peer networks
5. Data mining and analysis __ C. Optimizing business and marketing decisions