Suppose we are processing $100$ data items on a CPU cache. Each data item is $32$ bytes in size. And each I$/$O operation will fetch a $64$ byte cache line. Suppose the $100$ data items on the cache are contiguously stored.

In this case, is sequential reading faster than random reading? How much faster?

What if each data item is $64$ bytes in size instead of $32$ bytes? Seek time can be ignored for CPU cache.