NVIDIA interview question

Describe a memory access pattern that never misses in the cache

Interview Answers

Anonymous

22 Apr 2021

Linear patterns generally don't miss in the cache because of hardware prefetchers

Anonymous

22 Sept 2017

Write miss. Because write miss is updated by the processor into the physical / main memory. During this phase, the processor has the option to write to it's cache to maintain a local copy for later reference due to temporal or spatial locality. Write miss is further bifurcated into - write allocate and write no allocate.