On Wed, 14 Dec 2011, Shaohua Li wrote: > if vast majority of allocation needs picking from partial list of node, > the list_lock will have contention too. But I'd say avoiding the slab > thrashing does increase fastpath. Right, that's why my 2009 patchset would attempt to grab the partial slab with the highest number of free objects to a certain threshold before falling back to others and it improved performance somewhat. This was with the per-node partial lists, however, and the slowpath has been significantly rewritten since then. -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/ Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>