[A] Suppose we have a distributed program that needs to send data of size (n) from...

Transcribed Image Text:

[A] Suppose we have a distributed program that needs to send data of size (n) from computer A to computer B. Suppose there are two approaches to transfer the data: First, computer A sends the data (of size n) in one message to computer B. Second, computer A sends the same data in (m) messages to computer B. Suppose the network latency is (/) and the network bandwidth is (b). Compare the two approaches by showing how much time one of them is faster than the other? Assume that a computer system has 64MBytes (16M=2^24) of main memory, 64KBytes (16K=2^14) of cache size and 64 Bytes of cache line. A main memory address contains "A1E3BCD2" data in a particular memory cell. Transfer this data to the cache in direct mapping, associative mapping and 8-way set- associative mapping algorithms (show all major steps on the other side and make plausible assumptions, if necessary, with full justifications) [A] Suppose we have a distributed program that needs to send data of size (n) from computer A to computer B. Suppose there are two approaches to transfer the data: First, computer A sends the data (of size n) in one message to computer B. Second, computer A sends the same data in (m) messages to computer B. Suppose the network latency is (/) and the network bandwidth is (b). Compare the two approaches by showing how much time one of them is faster than the other? Assume that a computer system has 64MBytes (16M=2^24) of main memory, 64KBytes (16K=2^14) of cache size and 64 Bytes of cache line. A main memory address contains "A1E3BCD2" data in a particular memory cell. Transfer this data to the cache in direct mapping, associative mapping and 8-way set- associative mapping algorithms (show all major steps on the other side and make plausible assumptions, if necessary, with full justifications)