Consider a scenario where a high-resolution image scan at instrument Site A needs to be processed at a remote supercomputer Site B. The raw image transfer time between Site A and Site B is 2 minutes, and the processed image transfer time from Site B for final image rendering at Site A is 15 seconds. Assume: (i) the image comprises of 512 grids and is processed at Site B with four heterogeneous cores labeled C1, C2, C3 and C4; (ii) cores C1 and C2 have the same speed, and core C3 runs twice as fast as core C1, and core C4 runs eight times faster than core C1; (iii) it takes 1 second for core 1 or C2 to compute an individual grid within the image; (iv) following the division of labor in four cores C1 (64 grids), C2 (64 grids), C3 (128 grids), C4 (256 grids).

a) Calculate the total execution time starting from the raw image transfer from Site A till the final image rendering at Site A. What is the processor utilization rate at Site B?

b) If the rendering thin-clients core C5 at Site A can be leveraged to replace the compute execution at C1 (i.e., C1 is not used for the processing anymore), what is the speedup factor? You can assume core C5 is twice as fast as core C1. You can also assume the new raw image transfer time is 90 seconds, and the processed image transfer time for final image rendering is 12 seconds.