On Fri, 2022-04-22 at 14:43 +0800, Aaron Lu wrote: > On Fri, Apr 22, 2022 at 02:27:45PM +0800, ying.huang@xxxxxxxxx wrote: > > On Fri, 2022-04-22 at 14:24 +0800, Aaron Lu wrote: > > > On Thu, Apr 21, 2022 at 04:34:09PM +0800, ying.huang@xxxxxxxxx wrote: > > > > On Thu, 2022-04-21 at 16:17 +0800, Aaron Lu wrote: > > > > > On Thu, Apr 21, 2022 at 03:49:21PM +0800, ying.huang@xxxxxxxxx wrote: > > > > > > ... ... > > > > > > > > > For swap-in latency, we can use pmbench, which can output latency > > > > > > information. > > > > > > > > > > > > > > > > OK, I'll give pmbench a run, thanks for the suggestion. > > > > > > > > Better to construct a senario with more swapin than swapout. For > > > > example, start a memory eater, then kill it later. > > > > > > What about vm-scalability/case-swapin? > > > https://git.kernel.org/pub/scm/linux/kernel/git/wfg/vm-scalability.git/tree/case-swapin > > > > > > I think you are pretty familiar with it but still: > > > 1) it starts $nr_task processes and each mmaps $size/$nr_task area and > > > then consumes the memory, after this, it waits for a signal; > > > 2) start another process to consume $size memory to push the memory in > > > step 1) to swap device; > > > 3) kick processes in step 1) to start accessing their memory, thus > > > trigger swapins. The metric of this testcase is the swapin throughput. > > > > > > I plan to restrict the cgroup's limit to $size. > > > > > > Considering there is only one NVMe drive attached to node 0, I will run > > > the test as described before: > > > 1) bind processes to run on node 0, allocate on node 1 to test the > > > performance when reclaimer's node id is the same as swap device's. > > > 2) bind processes to run on node 1, allocate on node 0 to test the > > > performance when page's node id is the same as swap device's. > > > > > > Ying and Yang, > > > > > > Let me know what you think about the case used and the way the test is > > > conducted. > > > > The test case looks good to me. And, do you have a way to measure swap > > in latency? Better to compare between enabling and disabling per-node > > By swap in latency, do you mean the time it takes for a fault that is > served by swap in? > > Since the test is swap in only, I think throughput can tell us the > average swap in latency? > Yes. Given the same parallel level, the average swap in latency can be reflect via throughput. > > swap device support too to make sure per-node support has performance > > impact on this system. > > I think we can tell by conducting two more tests: > 1) bind processes to run on node 0, allocate on node 0; > 2) bind processes to run on node 1, allocate on node 1. > If case 1) is faster, we can say per-node support has performance impact > on this system. At least we can measure whether cross-node latency matters with this test. Best Regards, Huang, Ying
