3 ms·Preserving the Virtual Memory Abstraction15 points by jcbeard 9y agoanalyst_9 9y agoI've been hearing this from multiple places. Seems obvious after some talks and HN articles.anns 9y agointeresting. i knee this was an issue for accelerators but haven’t really thought about regular cores being bottlenecked by the TLB.