Re: Chelsio Forwarding performance and RELENG_13 vs RELENG_12
Date: Fri, 21 Oct 2022 18:45:37 UTC
On 10/21/2022 2:30 PM, Navdeep Parhar wrote: > On 10/21/22 11:15 AM, mike tancsa wrote: >> On 10/21/2022 2:13 PM, Navdeep Parhar wrote: >>> >>>>> hw.cxgbe.ntxq=4 >>>>> hw.cxgbe.nrxq=4 >>>>> >>>> Thanks Navdeep! >>>> >>>> Unfortunately, still the odd dropped packet :( >>> >>> Can you try increasing the size of the queues? >>> >>> hw.cxgbe.qsize_txq=2048 >>> hw.cxgbe.qsize_rxq=2048 >>> >>> The stats show that you are using MTU 1500. If you were using MTU >>> 9000 I'd also have suggested setting largest_rx_cluster to 4K. >>> >>> hw.cxgbe.largest_rx_cluster=4096 >>> >> Thanks again, just 1500 MTU. Shall I keep nt and nrxq at 4 still ? > > Yes, I think 4 queues are enough for 10G. > Sadly, no luck. Still about the same rate of overflows :( dev.cxl.0.stats.rx_trunc3: 0 dev.cxl.0.stats.rx_trunc2: 0 dev.cxl.0.stats.rx_trunc1: 0 dev.cxl.0.stats.rx_trunc0: 5 dev.cxl.0.stats.rx_ovflow3: 0 dev.cxl.0.stats.rx_ovflow2: 0 dev.cxl.0.stats.rx_ovflow1: 0 dev.cxl.0.stats.rx_ovflow0: 61 dev.cxl.0.stats.rx_ppp7: 0 dev.cxl.0.stats.rx_ppp6: 0 dev.cxl.0.stats.rx_ppp5: 0 dev.cxl.0.stats.rx_ppp4: 0 dev.cxl.0.stats.rx_ppp3: 0 dev.cxl.0.stats.rx_ppp2: 0 dev.cxl.0.stats.rx_ppp1: 0 dev.cxl.0.stats.rx_ppp0: 0 dev.cxl.0.stats.rx_pause: 0 dev.cxl.0.stats.rx_frames_1519_max: 0 dev.cxl.0.stats.rx_frames_1024_1518: 25966401 dev.cxl.0.stats.rx_frames_512_1023: 1569927 dev.cxl.0.stats.rx_frames_256_511: 1856460 dev.cxl.0.stats.rx_frames_128_255: 2718106 dev.cxl.0.stats.rx_frames_65_127: 4548626 dev.cxl.0.stats.rx_frames_64: 3044300 dev.cxl.0.stats.rx_runt: 0 dev.cxl.0.stats.rx_symbol_err: 0 dev.cxl.0.stats.rx_len_err: 0 dev.cxl.0.stats.rx_fcs_err: 0 dev.cxl.0.stats.rx_jabber: 0 dev.cxl.0.stats.rx_too_long: 0 dev.cxl.0.stats.rx_ucast_frames: 39618919 dev.cxl.0.stats.rx_mcast_frames: 36716 dev.cxl.0.stats.rx_bcast_frames: 48219 dev.cxl.0.stats.rx_frames: 39703866 dev.cxl.0.stats.rx_octets: 39678756369 dev.cxl.0.stats.tx_ppp7: 0