Question: An 8-core multi-core processor has an ideal speedup for embarrassingly parallel code of 8. The measured execution time of a benchmark code using only a

An 8-core multi-core processor has an ideal speedup for embarrassingly parallel code of 8. The measured execution time of a benchmark code using only a single thread (1), is 100 microseconds. The threading overhead needed to use OpenMP on the processor with 8 threads is 10 microseconds Taking the threading overhead of OpenMP, what is the expected, actual measured speed observed when the benchmark code is run on the processor with 8 threads? Assuming the same multi core processor and benchmark code What is the expected, actual speed up (accounting for threading overhead), if the overhead time is 1 microsecond
Step by Step Solution
There are 3 Steps involved in it
Get step-by-step solutions from verified subject matter experts
