// We have optimized the heap sizes to reduce TLB misses. Change this and you could &%$# up performance (by as much as 1+ ms/frame) Cache misses are one thing

// We have optimized the heap sizes to reduce TLB misses. Change this and you could &%$# up performance (by as much as 1+ ms/frame) Cache misses are one thing but here it's virtual memory and page tables!? TLB = Translation Lookaside Buffer