Question

Course Name: Cloud Computing

Is it ever possible to achieve super-linear speedup in parallel computing? If yes, give an example. If no, explain why not.

Answer 1

Hope this will clear your doubts .please let me know if you have any queries in the comment box

- Super linear speed up comes from exceeding neively calculated speedup even after taking into t the communication Process (which is faling, but still this is the bottle neck). Ego You have serial algorithm that takes it to execute you have tozy coles, so naive speed up is 1024x, kit takes t/1024, but it should be like in cor equation taking account menoly transfer slight modifications to the algorithm , Parallelism time so speed up should be lower than tozyo, but some times it happens that speedup is bigger, then we call it superlinear. - S In short superlinear speed up is achieved inthen the total amount of work processors do is stritaly less than the total work performes by a single processor. This can happen in three ways.

- Super linear speed up comes from exceeding neively calculated speedup even after taking into t the communication Process (which is faling, but still this is the bottle neck). Ego You have serial algorithm that takes it to execute you have tozy coles, so naive speed up is 1024x, kit takes t/1024, but it should be like in cor equation taking account menoly transfer slight modifications to the algorithm , Parallelism time so speed up should be lower than tozyo, but some times it happens that speedup is bigger, then we call it superlinear. - S In short superlinear speed up is achieved inthen the total amount of work processors do is stritaly less than the total work performes by a single processor. This can happen in three ways.

The original sequential algorithm was really bad, wing the parallel version of the algorithm on one Processor will usually do away with the super linear speedup. The parallel algorithm cases some search like arandom Walk, the more Processors that are walking, the less distance has to be walked in total before you reach what you looking tool for. Modern Processols have faster and slowe memories Typically it will try to keep the data you are using in the post memory. We can safely say I oor amount of data is larger than the amout of fast memory. If you use in processors you haven times the amount of faster memory. More data fits in fast memory which makes it possible to takes textime (this amant of work) on the Same task