On Monday 02 April 2007 17:37, Christoph Lameter wrote:On Sun, 1 Apr 2007, Andi Kleen wrote:
Hmm, this means there is at least 2MB worth of struct page on every node?Correct. 2MB worth of struct page is 128 mb of memory. Are there nodes with smaller amounts of memory?
Or do you have overlaps with other memory (I think you have)
In that case you have to handle the overlap in change_page_attr()
Yes the discontigmem minimum is 64MB and there are some setups
(mostly with numa emulation) where you end up with nodes that small.
Do you have any benchmarks numbers to prove it? There seem to be a fewYou want a benchmark to prove that the removal of memory references and code improves performance?
benchmarks where the discontig virt_to_page is a problem
(although I know ways to make it more efficient), and sparsemem
is normally slower. Still some numbers would be good.
You're just moving them into MMU, not really removing it. And need more TLB entries.
It might be faster or it might not. There are some unexpected issues, like most x86-64 CPUs have a quite small number of large TLBs so you can get thrashing etc.
So numbers with TLB intensive workloads would be good.