Hi Raj, > On Jan 30, 2019, at 6:49 AM, Raj <rajlistu...@gmail.com> wrote: > > Hello Florin, > > Thanks a lot for your reply. > > On Tue, Jan 29, 2019 at 9:15 PM Florin Coras <fcoras.li...@gmail.com> wrote: >> >> VPP typically works best if configured with a relatively small number of >> buffers. > > I am trying to understand why this would be the case.
Because if it can work with all of its buffers cached, without flushing them to main memory, it can access and therefore ‘process’ them faster. > >> On the other hand, nsim can induce a large amount of delay, so the buffers >> cannot be used to temporarily store the data. > > If my understanding is correct, the comparison is between two packet > copy per packet and performance degradation due to bigger buffers. How > would they compare? It's two copies vs more buffer memory, not bigger buffers. Flushing and reading buffers to/from main memory are pretty much copy operations. Nsim typically needs a lot of buffer memory, therefore instead of trashing the cache and potentially impacting other vpp components, it directly stores everything in memory. Florin > > Thanks and Regards, > > Raj > -=-=-=-=-=-=-=-=-=-=-=- > Links: You receive all messages sent to this group. > > View/Reply Online (#12065): https://lists.fd.io/g/vpp-dev/message/12065 > Mute This Topic: https://lists.fd.io/mt/29581029/675152 > Group Owner: vpp-dev+ow...@lists.fd.io > Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [fcoras.li...@gmail.com] > -=-=-=-=-=-=-=-=-=-=-=-
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#12068): https://lists.fd.io/g/vpp-dev/message/12068 Mute This Topic: https://lists.fd.io/mt/29581029/21656 Group Owner: vpp-dev+ow...@lists.fd.io Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
