Tighter bitcell pitch required for L4 cache compared to the eNVM application