CareerCup

Probably locality of reference.

is performance better for searching an element?

Binary tree can allocate memory from anywhere in the heap...these locations need not be on the same page...in an array they are contiguos memory and hence once the page is swapped in it stays in primary memory and all access are fast...similar to link list vs array

Less cache misses for array because of its better locality of reference.

I think frequent updation in binary tree will cause fragmentation, which is not the case with arrays

First of all, I am not sure if his assertion is correct.
Event if it is correct, I doubt it has to do with computer architecture. It might has to do with the size of memory. It might take longer to address the memory in a 4Gb memory than a 4kb memory.
I might be wrong.

@XYZ: "I might be wrong......" yeh baby you're completely wrong !

its prefetching in case of array.

200-elements array can be kept in L1 cache (which is effectively the fastest computer memory) entirely, while 200-elements tree with high probability can not be kept there (a tree is fragmented beacuse of memory block for each node is usually allocated separetely from the heap).