On Mon, Jun 13, 2016 at 11:50:48AM -0700, Javier Blazquez wrote: >[...] > This is a very exciting patch. I applied it and reran some microbenchmarks > of mine that test the TX and RX paths separately. These are the results I > got: > > TX path (burst = 64 packets) > > 1 thread - 2 ports - 4 queues per port: 39Mpps => 48Mpps > 2 threads - 2 ports - 2 queues per port: 60Mpps => 60Mpps (hardware > limitation?)
To be able to reach higher values you will need to configure the inline feature with the device argument txq_inline, and only activate it with more than 1 queue, this can be done with the txq_min_inline argument. This feature helps the NIC by reducing the PCI back-pressure, in counterpart it will consume more CPU cycles. You can take a look to the NIC documentation (doc/guides/nics/mlx5.rst) updated in this path-set which explains both txq_inline and txqs_min_inline device arguments. > RX path (burst = 32 packets) > > 1 thread - 2 ports - 4 queues per port: 38Mpps => 46Mpps > 2 threads - 2 ports - 2 queues per port: 43Mpps => 50Mpps > > The tests were run on the following hardware, using DPDK master with this > patch and the "Miscellaneous fixes for mlx4 and mlx5" patch applied: > > 2x Intel Xeon E5-2680 v3 2.5GHz > 64GB DDR4-2133 > 1x Mellanox ConnectX-4 EN, 40/56GbE dual-port, PCIe3.0 x8 (MCX414A-BCAT) > > I haven't test it extensively outside of these microbenchmarks, but so far > this patch has been working great on my end, so: > > tested-by: Javier Blazquez <jblazquez at riotgames.com> Regards, -- N?lio Laranjeiro 6WIND
