Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser
Hello Yichen and Steven 1. Steven- I will upgrade qemu to 2.8 and let you know the perf. results. 2. Yichen - I took ubuntu 16.04 iso and launch its qcow2 like this: sudo qemu-system-x86_64 -name vhost1 -m 2048M -smp 4 -cpu host -hda /var/lib/libvirt/images/vm1.qcow2 -boot c -enable-kvm -no-reboot -net none -chardev socket,id=char1,path=/var/run/vpp/sock2.sock -netdev type=vhost-user,id=mynet1,chardev=char1,vhostforce -device virtio-net-pci,mac=c0:00:00:00:00:ff,netdev=mynet1 -object memory-backend-file,id=mem,size=2048M,mem-path=/dev/hugepages,share=on -numa node,memdev=mem -mem-prealloc - other steps are like in https://wiki.fd.io/view/VPP/Use_VPP_to_connect_VMs_Using_Vhost-User_Interface i get pps performance like in ovs-dpdk, regarding the latency - when i ping i get ~100us latency , but stiil when i ping -f iget almost 1ms latency Maybe it has to do with MAX_BURST_SIZE (=256 ) because ping -f has relatevely high rate and the dpdk fills buffer with 256 pkts , unlike the regular ping with low rate - there the DPDK exit the loop *before* filling the buffer with 256 pkts ... -Sara On Mon, Mar 26, 2018 at 6:48 AM, steven luong wrote: > Sara, > > > > You need qemu 2.8.0 or later for the rx_queue_size option to work. > > > > Steven > > > > *From: * on behalf of Sara Gittlin < > sara.gitt...@gmail.com> > *Date: *Sunday, March 25, 2018 at 1:57 AM > > *To: *"vpp-dev@lists.fd.io" > *Cc: *"vpp-dev@lists.fd.io" > *Subject: *Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser > > > > Thank Steven and Alec > > I'm using the iperf as a basic tool to check sanity - i'll run pktgen on > the VMs later. > > the resut i'm got 20Mbps (~20K pps) throughput and latency of 1ms is not > iperf issue. > > Steven - you suggest to set the rx-queue-size when lanuching the VM but > unfortunately i dont find this option in the qemy-system-x86_64 command line > > i'm using this: > > qemu-system-x86_64 \ > > -enable-kvm -m 1024 \ > > -bios OVMF.fd \ > > -smp 4 -cpu host \ > > -vga none -nographic \ > > -drive file="1-clear-14200-kvm.img",if=virtio,aio=threads \ > > -chardev socket,id=char1,path=/var/run/vpp/sock1.sock \ > > -netdev type=vhost-user,id=mynet1,chardev=char1,vhostforce \ > > -device virtio-net-pci,mac=00:00:00:00:00:01,netdev=mynet1 \ > > -object > memory-backend-file,id=mem,size=1024M,mem-path=/dev/hugepages,share=on \ > > -numa node,memdev=mem -mem-prealloc \ > > -debugcon file:debug.log -global isa-debugcon.iobase=0x402 > > > > -Sara > > > > > > On Thu, Mar 22, 2018 at 6:10 PM, steven luong wrote: > > Sara, > > > > Iperf3 is not blasting traffic fast enough. You could try specifying > multiple parallel streams using -P. Then, you will likely encounter > vhost-user dropping packets as you are using the default virtqueue size > 256. You’ll need to specify rx_queue_size=1024 when you launch the VM in > the qemu start command to reduce the drops. > > > > Steven > > > > *From: * on behalf of Sara Gittlin < > sara.gitt...@gmail.com> > *Date: *Thursday, March 22, 2018 at 8:12 AM > *To: *"vpp-dev@lists.fd.io" > *Cc: *"vpp-dev@lists.fd.io" > > > *Subject: *Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser > > > > Hi John and Steven > > > > setting this in the startup config didnt help > > vhost-user { > > coalesce-frame 0 > > } > > John > > > > I'm using ping -f for latency and iperf3 for pps testing. > > later i'll run pktgen in the vm's > > > > output : > > sho int > Name Idx State Counter > Count > VirtualEthernet0/0/0 1 up rx > packets 11748 > rx > bytes 868648 > tx > packets 40958 > tx > bytes58047352 > > drops 29 > VirtualEthernet0/0/1 2 up rx > packets 40958 > rx > bytes58047352 > tx > packets 11719 > tx > bytes 862806 > > tx-error 29 > local00 up > > > > > > show h
Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser
Sara, You need qemu 2.8.0 or later for the rx_queue_size option to work. Steven From: on behalf of Sara Gittlin Date: Sunday, March 25, 2018 at 1:57 AM To: "vpp-dev@lists.fd.io" Cc: "vpp-dev@lists.fd.io" Subject: Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser Thank Steven and Alec I'm using the iperf as a basic tool to check sanity - i'll run pktgen on the VMs later. the resut i'm got 20Mbps (~20K pps) throughput and latency of 1ms is not iperf issue. Steven - you suggest to set the rx-queue-size when lanuching the VM but unfortunately i dont find this option in the qemy-system-x86_64 command line i'm using this: qemu-system-x86_64 \ -enable-kvm -m 1024 \ -bios OVMF.fd \ -smp 4 -cpu host \ -vga none -nographic \ -drive file="1-clear-14200-kvm.img",if=virtio,aio=threads \ -chardev socket,id=char1,path=/var/run/vpp/sock1.sock \ -netdev type=vhost-user,id=mynet1,chardev=char1,vhostforce \ -device virtio-net-pci,mac=00:00:00:00:00:01,netdev=mynet1 \ -object memory-backend-file,id=mem,size=1024M,mem-path=/dev/hugepages,share=on \ -numa node,memdev=mem -mem-prealloc \ -debugcon file:debug.log -global isa-debugcon.iobase=0x402 -Sara On Thu, Mar 22, 2018 at 6:10 PM, steven luong mailto:slu...@cisco.com>> wrote: Sara, Iperf3 is not blasting traffic fast enough. You could try specifying multiple parallel streams using -P. Then, you will likely encounter vhost-user dropping packets as you are using the default virtqueue size 256. You’ll need to specify rx_queue_size=1024 when you launch the VM in the qemu start command to reduce the drops. Steven From: mailto:vpp-dev@lists.fd.io>> on behalf of Sara Gittlin mailto:sara.gitt...@gmail.com>> Date: Thursday, March 22, 2018 at 8:12 AM To: "vpp-dev@lists.fd.io<mailto:vpp-dev@lists.fd.io>" mailto:vpp-dev@lists.fd.io>> Cc: "vpp-dev@lists.fd.io<mailto:vpp-dev@lists.fd.io>" mailto:vpp-dev@lists.fd.io>> Subject: Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser Hi John and Steven setting this in the startup config didnt help vhost-user { coalesce-frame 0 } John I'm using ping -f for latency and iperf3 for pps testing. later i'll run pktgen in the vm's output : sho int Name Idx State Counter Count VirtualEthernet0/0/0 1 up rx packets 11748 rx bytes 868648 tx packets 40958 tx bytes 58047352 drops 29 VirtualEthernet0/0/1 2 up rx packets 40958 rx bytes 58047352 tx packets 11719 tx bytes 862806 tx-error 29 local00 up show hardware NameIdx Link Hardware VirtualEthernet0/0/0 1 up VirtualEthernet0/0/0 Ethernet address 02:fe:25:2f:bd:c2 VirtualEthernet0/0/1 2 up VirtualEthernet0/0/1 Ethernet address 02:fe:40:16:70:1b local0 0down local0 local -Sara On Thu, Mar 22, 2018 at 4:53 PM, John DeNisco mailto:jdeni...@cisco.com>> wrote: Hi Sara, Can you also send the results from show hardware and show interfaces? What are you using to test your performance. John From: mailto:vpp-dev@lists.fd.io>> on behalf of Sara Gittlin mailto:sara.gitt...@gmail.com>> Date: Thursday, March 22, 2018 at 9:27 AM To: "vpp-dev@lists.fd.io<mailto:vpp-dev@lists.fd.io>" mailto:vpp-dev@lists.fd.io>> Subject: Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser this is the output of: show vhost-user VirtualEthernet0/0/0 Virtio vhost-user interfaces Global: coalesce frames 32 time 1e-3 number of rx virtqueues in interrupt mode: 0 Interface: VirtualEthernet0/0/0 (ifindex 1) virtio_net_hdr_sz 12 features mask (0x): features (0x58208000): VIRTIO_NET_F_MRG_RXBUF (15) VIRTIO_NET_F_GUEST_ANNOUNCE (21) VIRTIO_F_ANY_LAYOUT (27) VIRTIO_F_INDIRECT_DESC (28) VHOST_USER_F_PROTOCOL_FEATURES (30) protocol features (0x3) VHOST_USER_PROTOCOL_F_MQ (0) VHOST_USER_PROTOCOL_F_LOG_SHMFD (1) socket filename /var/run/vpp/sock1.sock type server errno "Success" rx placement: thread 1 on vring 1, polling tx placem
Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser
Hi, Sara, May I know what image you are using to achieving the good performance with VPP? I assume you are running the same iperf3 testings? Thanks very much! Regards, Yichen From: on behalf of Sara Gittlin Date: Sunday, March 25, 2018 at 5:31 AM To: "vpp-dev@lists.fd.io" Subject: Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser Fyi When i launch another image (not based on the clear-linux-kvm) - i'm achieving good performance (exactly like ovs-dpdk - which is expected in single flow/mac ) any idea why ? this is the launch cmdline: [anyway i cant find a way to set the tx/rx-queue-size) sudo qemu-system-x86_64 -name vhost1 -m 2048M -smp 4 -cpu host -hda /var/lib/libvirt/images/vm1.qcow2 -boot c -enable-kvm -no-reboot -net none -chardev socket,id=char3,path=/var/run/vpp/sock2.sock -netdev type=vhost-user,id=mynet1,chardev=char3,vhostforce -device virtio-net-pci,mac=c0:00:00:00:00:ff,netdev=mynet1 -object memory-backend-file,id=mem,size=2048M,mem-path=/dev/hugepages,share=on -numa node,memdev=mem -mem-prealloc Thank you -Sara On Sun, Mar 25, 2018 at 11:56 AM, Sara Gittlin mailto:sara.gitt...@gmail.com>> wrote: Thank Steven and Alec I'm using the iperf as a basic tool to check sanity - i'll run pktgen on the VMs later. the resut i'm got 20Mbps (~20K pps) throughput and latency of 1ms is not iperf issue. Steven - you suggest to set the rx-queue-size when lanuching the VM but unfortunately i dont find this option in the qemy-system-x86_64 command line i'm using this: qemu-system-x86_64 \ -enable-kvm -m 1024 \ -bios OVMF.fd \ -smp 4 -cpu host \ -vga none -nographic \ -drive file="1-clear-14200-kvm.img",if=virtio,aio=threads \ -chardev socket,id=char1,path=/var/run/vpp/sock1.sock \ -netdev type=vhost-user,id=mynet1,chardev=char1,vhostforce \ -device virtio-net-pci,mac=00:00:00:00:00:01,netdev=mynet1 \ -object memory-backend-file,id=mem,size=1024M,mem-path=/dev/hugepages,share=on \ -numa node,memdev=mem -mem-prealloc \ -debugcon file:debug.log -global isa-debugcon.iobase=0x402 -Sara On Thu, Mar 22, 2018 at 6:10 PM, steven luong mailto:slu...@cisco.com>> wrote: Sara, Iperf3 is not blasting traffic fast enough. You could try specifying multiple parallel streams using -P. Then, you will likely encounter vhost-user dropping packets as you are using the default virtqueue size 256. You’ll need to specify rx_queue_size=1024 when you launch the VM in the qemu start command to reduce the drops. Steven From: mailto:vpp-dev@lists.fd.io>> on behalf of Sara Gittlin mailto:sara.gitt...@gmail.com>> Date: Thursday, March 22, 2018 at 8:12 AM To: "vpp-dev@lists.fd.io<mailto:vpp-dev@lists.fd.io>" mailto:vpp-dev@lists.fd.io>> Cc: "vpp-dev@lists.fd.io<mailto:vpp-dev@lists.fd.io>" mailto:vpp-dev@lists.fd.io>> Subject: Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser Hi John and Steven setting this in the startup config didnt help vhost-user { coalesce-frame 0 } John I'm using ping -f for latency and iperf3 for pps testing. later i'll run pktgen in the vm's output : sho int Name Idx State Counter Count VirtualEthernet0/0/0 1 up rx packets 11748 rx bytes 868648 tx packets 40958 tx bytes 58047352 drops 29 VirtualEthernet0/0/1 2 up rx packets 40958 rx bytes 58047352 tx packets 11719 tx bytes 862806 tx-error 29 local00 up show hardware NameIdx Link Hardware VirtualEthernet0/0/0 1 up VirtualEthernet0/0/0 Ethernet address 02:fe:25:2f:bd:c2 VirtualEthernet0/0/1 2 up VirtualEthernet0/0/1 Ethernet address 02:fe:40:16:70:1b local0 0down local0 local -Sara On Thu, Mar 22, 2018 at 4:53 PM, John DeNisco mailto:jdeni...@cisco.com>> wrote: Hi Sara, Can you also send the results from show hardware and show interfaces? What are you using to test your performance. John From: mailto:vpp-dev@lists.fd.io>> on behalf of Sara Gittlin mailto:sara.gitt...@gmail.com>> Date: Thursday, March 22, 2018 at 9:27 AM To: "vpp-d
Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser
Fyi When i launch another image (not based on the clear-linux-kvm) - i'm achieving good performance (exactly like ovs-dpdk - which is expected in single flow/mac ) any idea why ? this is the launch cmdline: [anyway i cant find a way to set the tx/rx-queue-size) sudo qemu-system-x86_64 -name vhost1 -m 2048M -smp 4 -cpu host -hda /var/lib/libvirt/images/vm1.qcow2 -boot c -enable-kvm -no-reboot -net none -chardev socket,id=char3,path=/var/run/vpp/sock2.sock -netdev type=vhost-user,id=mynet1,chardev=char3,vhostforce -device virtio-net-pci,mac=c0:00:00:00:00:ff,netdev=mynet1 -object memory-backend-file,id=mem,size=2048M,mem-path=/dev/hugepages,share=on -numa node,memdev=mem -mem-prealloc Thank you -Sara On Sun, Mar 25, 2018 at 11:56 AM, Sara Gittlin wrote: > Thank Steven and Alec > I'm using the iperf as a basic tool to check sanity - i'll run pktgen on > the VMs later. > the resut i'm got 20Mbps (~20K pps) throughput and latency of 1ms is not > iperf issue. > Steven - you suggest to set the rx-queue-size when lanuching the VM but > unfortunately i dont find this option in the qemy-system-x86_64 command line > > i'm using this: > > qemu-system-x86_64 \ > -enable-kvm -m 1024 \ > -bios OVMF.fd \ > -smp 4 -cpu host \ > -vga none -nographic \ > -drive file="1-clear-14200-kvm.img",if=virtio,aio=threads \ > -chardev socket,id=char1,path=/var/run/vpp/sock1.sock \ > -netdev type=vhost-user,id=mynet1,chardev=char1,vhostforce \ > -device virtio-net-pci,mac=00:00:00:00:00:01,netdev=mynet1 \ > -object > memory-backend-file,id=mem,size=1024M,mem-path=/dev/hugepages,share=on \ > -numa node,memdev=mem -mem-prealloc \ > -debugcon file:debug.log -global isa-debugcon.iobase=0x402 > > -Sara > > > > On Thu, Mar 22, 2018 at 6:10 PM, steven luong wrote: > >> Sara, >> >> >> >> Iperf3 is not blasting traffic fast enough. You could try specifying >> multiple parallel streams using -P. Then, you will likely encounter >> vhost-user dropping packets as you are using the default virtqueue size >> 256. You’ll need to specify rx_queue_size=1024 when you launch the VM in >> the qemu start command to reduce the drops. >> >> >> >> Steven >> >> >> >> *From: * on behalf of Sara Gittlin < >> sara.gitt...@gmail.com> >> *Date: *Thursday, March 22, 2018 at 8:12 AM >> *To: *"vpp-dev@lists.fd.io" >> *Cc: *"vpp-dev@lists.fd.io" >> >> *Subject: *Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser >> >> >> >> Hi John and Steven >> >> >> >> setting this in the startup config didnt help >> >> vhost-user { >> >> coalesce-frame 0 >> >> } >> >> John >> >> >> >> I'm using ping -f for latency and iperf3 for pps testing. >> >> later i'll run pktgen in the vm's >> >> >> >> output : >> >> sho int >> Name Idx State >> Counter Count >> VirtualEthernet0/0/0 1 up rx >> packets 11748 >> rx >> bytes 868648 >> tx >> packets 40958 >> tx >> bytes58047352 >> >> drops 29 >> VirtualEthernet0/0/1 2 up rx >> packets 40958 >> rx >> bytes58047352 >> tx >> packets 11719 >> tx >> bytes 862806 >> >> tx-error 29 >> local00 up >> >> >> >> >> >> show hardware >> NameIdx Link Hardware >> VirtualEthernet0/0/0 1 up VirtualEthernet0/0/0 >> Ethernet address 02:fe:25:2f:bd:c2 >> VirtualEthernet0/0/1 2 up VirtualEthernet0/0/1 >> Ethernet address 02:fe:40:16:70:1b >> local0 0down local0 >> local >> >> >> >> -Sara >> >> >> >> >> >> On Thu, Mar 22, 2018 at 4:53 PM, John DeNisco wrote: >> >> >> >> Hi Sara, >> >> >> >>
Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser
Thank Steven and Alec I'm using the iperf as a basic tool to check sanity - i'll run pktgen on the VMs later. the resut i'm got 20Mbps (~20K pps) throughput and latency of 1ms is not iperf issue. Steven - you suggest to set the rx-queue-size when lanuching the VM but unfortunately i dont find this option in the qemy-system-x86_64 command line i'm using this: qemu-system-x86_64 \ -enable-kvm -m 1024 \ -bios OVMF.fd \ -smp 4 -cpu host \ -vga none -nographic \ -drive file="1-clear-14200-kvm.img",if=virtio,aio=threads \ -chardev socket,id=char1,path=/var/run/vpp/sock1.sock \ -netdev type=vhost-user,id=mynet1,chardev=char1,vhostforce \ -device virtio-net-pci,mac=00:00:00:00:00:01,netdev=mynet1 \ -object memory-backend-file,id=mem,size=1024M,mem-path=/dev/hugepages,share=on \ -numa node,memdev=mem -mem-prealloc \ -debugcon file:debug.log -global isa-debugcon.iobase=0x402 -Sara On Thu, Mar 22, 2018 at 6:10 PM, steven luong wrote: > Sara, > > > > Iperf3 is not blasting traffic fast enough. You could try specifying > multiple parallel streams using -P. Then, you will likely encounter > vhost-user dropping packets as you are using the default virtqueue size > 256. You’ll need to specify rx_queue_size=1024 when you launch the VM in > the qemu start command to reduce the drops. > > > > Steven > > > > *From: * on behalf of Sara Gittlin < > sara.gitt...@gmail.com> > *Date: *Thursday, March 22, 2018 at 8:12 AM > *To: *"vpp-dev@lists.fd.io" > *Cc: *"vpp-dev@lists.fd.io" > > *Subject: *Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser > > > > Hi John and Steven > > > > setting this in the startup config didnt help > > vhost-user { > > coalesce-frame 0 > > } > > John > > > > I'm using ping -f for latency and iperf3 for pps testing. > > later i'll run pktgen in the vm's > > > > output : > > sho int > Name Idx State Counter > Count > VirtualEthernet0/0/0 1 up rx > packets 11748 > rx > bytes 868648 > tx > packets 40958 > tx > bytes58047352 > > drops 29 > VirtualEthernet0/0/1 2 up rx > packets 40958 > rx > bytes58047352 > tx > packets 11719 > tx > bytes 862806 > > tx-error 29 > local00 up > > > > > > show hardware > NameIdx Link Hardware > VirtualEthernet0/0/0 1 up VirtualEthernet0/0/0 > Ethernet address 02:fe:25:2f:bd:c2 > VirtualEthernet0/0/1 2 up VirtualEthernet0/0/1 > Ethernet address 02:fe:40:16:70:1b > local0 0down local0 > local > > > > -Sara > > > > > > On Thu, Mar 22, 2018 at 4:53 PM, John DeNisco wrote: > > > > Hi Sara, > > > > Can you also send the results from show hardware and show interfaces? > > > > What are you using to test your performance. > > > > John > > > > > > *From: * on behalf of Sara Gittlin < > sara.gitt...@gmail.com> > *Date: *Thursday, March 22, 2018 at 9:27 AM > *To: *"vpp-dev@lists.fd.io" > *Subject: *Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser > > > > this is the output of: > > > show vhost-user VirtualEthernet0/0/0 > Virtio vhost-user interfaces > Global: > coalesce frames 32 time 1e-3 > number of rx virtqueues in interrupt mode: 0 > Interface: VirtualEthernet0/0/0 (ifindex 1) > virtio_net_hdr_sz 12 > features mask (0x): > features (0x58208000): >VIRTIO_NET_F_MRG_RXBUF (15) >VIRTIO_NET_F_GUEST_ANNOUNCE (21) >VIRTIO_F_ANY_LAYOUT (27) >VIRTIO_F_INDIRECT_DESC (28) >VHOST_USER_F_PROTOCOL_FEATURES (30) > protocol features (0x3) >VHOST_USER_PROTOCOL_F_MQ (0) >VHOST_USER_PROTOCOL_F_LOG_SHMFD (1) > > socket filename /var/run/vpp/sock1.sock type server errno "Success" > > rx placement: >thread 1 on vring 1, polling > tx placement: spin-lock >thread 0 on vring 0 >thread 1
Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser
Sara, Iperf3 is not blasting traffic fast enough. You could try specifying multiple parallel streams using -P. Then, you will likely encounter vhost-user dropping packets as you are using the default virtqueue size 256. You’ll need to specify rx_queue_size=1024 when you launch the VM in the qemu start command to reduce the drops. Steven From: on behalf of Sara Gittlin Date: Thursday, March 22, 2018 at 8:12 AM To: "vpp-dev@lists.fd.io" Cc: "vpp-dev@lists.fd.io" Subject: Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser Hi John and Steven setting this in the startup config didnt help vhost-user { coalesce-frame 0 } John I'm using ping -f for latency and iperf3 for pps testing. later i'll run pktgen in the vm's output : sho int Name Idx State Counter Count VirtualEthernet0/0/0 1 up rx packets 11748 rx bytes 868648 tx packets 40958 tx bytes 58047352 drops 29 VirtualEthernet0/0/1 2 up rx packets 40958 rx bytes 58047352 tx packets 11719 tx bytes 862806 tx-error 29 local00 up show hardware NameIdx Link Hardware VirtualEthernet0/0/0 1 up VirtualEthernet0/0/0 Ethernet address 02:fe:25:2f:bd:c2 VirtualEthernet0/0/1 2 up VirtualEthernet0/0/1 Ethernet address 02:fe:40:16:70:1b local0 0down local0 local -Sara On Thu, Mar 22, 2018 at 4:53 PM, John DeNisco mailto:jdeni...@cisco.com>> wrote: Hi Sara, Can you also send the results from show hardware and show interfaces? What are you using to test your performance. John From: mailto:vpp-dev@lists.fd.io>> on behalf of Sara Gittlin mailto:sara.gitt...@gmail.com>> Date: Thursday, March 22, 2018 at 9:27 AM To: "vpp-dev@lists.fd.io<mailto:vpp-dev@lists.fd.io>" mailto:vpp-dev@lists.fd.io>> Subject: Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser this is the output of: show vhost-user VirtualEthernet0/0/0 Virtio vhost-user interfaces Global: coalesce frames 32 time 1e-3 number of rx virtqueues in interrupt mode: 0 Interface: VirtualEthernet0/0/0 (ifindex 1) virtio_net_hdr_sz 12 features mask (0x): features (0x58208000): VIRTIO_NET_F_MRG_RXBUF (15) VIRTIO_NET_F_GUEST_ANNOUNCE (21) VIRTIO_F_ANY_LAYOUT (27) VIRTIO_F_INDIRECT_DESC (28) VHOST_USER_F_PROTOCOL_FEATURES (30) protocol features (0x3) VHOST_USER_PROTOCOL_F_MQ (0) VHOST_USER_PROTOCOL_F_LOG_SHMFD (1) socket filename /var/run/vpp/sock1.sock type server errno "Success" rx placement: thread 1 on vring 1, polling tx placement: spin-lock thread 0 on vring 0 thread 1 on vring 0 thread 2 on vring 0 Memory regions (total 2) region fdguest_phys_addrmemory_sizeuserspace_addr mmap_offsetmmap_addr == = == == == == == 0 320x 0x000c 0x7f5a6a60 0x 0x7f47c420 1 330x0010 0x3ff0 0x7f5a6a70 0x0010 0x7f46f410 Virtqueue 0 (TX) qsz 256 last_avail_idx 63392 last_used_idx 63392 avail.flags 0 avail.idx 63530 used.flags 1 used.idx 63392 kickfd 34 callfd 35 errfd -1 Virtqueue 1 (RX) qsz 256 last_avail_idx 32414 last_used_idx 32414 avail.flags 1 avail.idx 32414 used.flags 1 used.idx 32414 kickfd 30 callfd 36 errfd -1 On Thu, Mar 22, 2018 at 3:07 PM, Sara Gittlin mailto:sara.gitt...@gmail.com>> wrote: i dont think these are error counters - anyway very poor pps On Thu, Mar 22, 2018 at 2:55 PM, Sara Gittlin mailto:sara.gitt...@gmail.com>> wrote: in the show err output i see that l2-output l2-learn l2-input counters are continuously incremented : show err CountNode Reason 11l2-output L2 output packets 11l2-learnL2 learn packets 11l2-inputL2 input packets 3l2-floodL2 flood packets 8479644l2-output
Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser
You should not use iperf for pps because you’re going to incur the overhead of running a guest user space app over socket + guest kernel tcp-ip stack + virtio over vhost user (times 2 since both your end points run iperf). You might get decent throughput with large write buffer size (e.g. 64k writes) as those do not require as many pps. For pps in VM it is better to use pktgen or any equivalent traffic generator based on DPDK. Alec From: on behalf of Sara Gittlin Date: Thursday, March 22, 2018 at 8:12 AM To: "vpp-dev@lists.fd.io" Cc: "vpp-dev@lists.fd.io" Subject: Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser Hi John and Steven setting this in the startup config didnt help vhost-user { coalesce-frame 0 } John I'm using ping -f for latency and iperf3 for pps testing. later i'll run pktgen in the vm's output : sho int Name Idx State Counter Count VirtualEthernet0/0/0 1 up rx packets 11748 rx bytes 868648 tx packets 40958 tx bytes 58047352 drops 29 VirtualEthernet0/0/1 2 up rx packets 40958 rx bytes 58047352 tx packets 11719 tx bytes 862806 tx-error 29 local00 up show hardware NameIdx Link Hardware VirtualEthernet0/0/0 1 up VirtualEthernet0/0/0 Ethernet address 02:fe:25:2f:bd:c2 VirtualEthernet0/0/1 2 up VirtualEthernet0/0/1 Ethernet address 02:fe:40:16:70:1b local0 0down local0 local -Sara On Thu, Mar 22, 2018 at 4:53 PM, John DeNisco mailto:jdeni...@cisco.com>> wrote: Hi Sara, Can you also send the results from show hardware and show interfaces? What are you using to test your performance. John From: mailto:vpp-dev@lists.fd.io>> on behalf of Sara Gittlin mailto:sara.gitt...@gmail.com>> Date: Thursday, March 22, 2018 at 9:27 AM To: "vpp-dev@lists.fd.io<mailto:vpp-dev@lists.fd.io>" mailto:vpp-dev@lists.fd.io>> Subject: Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser this is the output of: show vhost-user VirtualEthernet0/0/0 Virtio vhost-user interfaces Global: coalesce frames 32 time 1e-3 number of rx virtqueues in interrupt mode: 0 Interface: VirtualEthernet0/0/0 (ifindex 1) virtio_net_hdr_sz 12 features mask (0x): features (0x58208000): VIRTIO_NET_F_MRG_RXBUF (15) VIRTIO_NET_F_GUEST_ANNOUNCE (21) VIRTIO_F_ANY_LAYOUT (27) VIRTIO_F_INDIRECT_DESC (28) VHOST_USER_F_PROTOCOL_FEATURES (30) protocol features (0x3) VHOST_USER_PROTOCOL_F_MQ (0) VHOST_USER_PROTOCOL_F_LOG_SHMFD (1) socket filename /var/run/vpp/sock1.sock type server errno "Success" rx placement: thread 1 on vring 1, polling tx placement: spin-lock thread 0 on vring 0 thread 1 on vring 0 thread 2 on vring 0 Memory regions (total 2) region fdguest_phys_addrmemory_sizeuserspace_addr mmap_offsetmmap_addr == = == == == == == 0 320x 0x000c 0x7f5a6a60 0x 0x7f47c420 1 330x0010 0x3ff0 0x7f5a6a70 0x0010 0x7f46f410 Virtqueue 0 (TX) qsz 256 last_avail_idx 63392 last_used_idx 63392 avail.flags 0 avail.idx 63530 used.flags 1 used.idx 63392 kickfd 34 callfd 35 errfd -1 Virtqueue 1 (RX) qsz 256 last_avail_idx 32414 last_used_idx 32414 avail.flags 1 avail.idx 32414 used.flags 1 used.idx 32414 kickfd 30 callfd 36 errfd -1 On Thu, Mar 22, 2018 at 3:07 PM, Sara Gittlin mailto:sara.gitt...@gmail.com>> wrote: i dont think these are error counters - anyway very poor pps On Thu, Mar 22, 2018 at 2:55 PM, Sara Gittlin mailto:sara.gitt...@gmail.com>> wrote: in the show err output i see that l2-output l2-learn l2-input counters are continuously incremented : show err CountNode Reason 11l2-output L2 output packets 11l2-learnL2 learn packets 11l2-inputL2 input packets
Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser
Hi John and Steven setting this in the startup config didnt help vhost-user { coalesce-frame 0 } John I'm using ping -f for latency and iperf3 for pps testing. later i'll run pktgen in the vm's output : sho int Name Idx State Counter Count VirtualEthernet0/0/0 1 up rx packets 11748 rx bytes 868648 tx packets 40958 tx bytes58047352 drops 29 VirtualEthernet0/0/1 2 up rx packets 40958 rx bytes58047352 tx packets 11719 tx bytes 862806 tx-error 29 local00 up show hardware NameIdx Link Hardware VirtualEthernet0/0/0 1 up VirtualEthernet0/0/0 Ethernet address 02:fe:25:2f:bd:c2 VirtualEthernet0/0/1 2 up VirtualEthernet0/0/1 Ethernet address 02:fe:40:16:70:1b local0 0down local0 local -Sara On Thu, Mar 22, 2018 at 4:53 PM, John DeNisco wrote: > > > Hi Sara, > > > > Can you also send the results from show hardware and show interfaces? > > > > What are you using to test your performance. > > > > John > > > > > > *From: * on behalf of Sara Gittlin < > sara.gitt...@gmail.com> > *Date: *Thursday, March 22, 2018 at 9:27 AM > *To: *"vpp-dev@lists.fd.io" > *Subject: *Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser > > > > this is the output of: > > show vhost-user VirtualEthernet0/0/0 > Virtio vhost-user interfaces > Global: > coalesce frames 32 time 1e-3 > number of rx virtqueues in interrupt mode: 0 > Interface: VirtualEthernet0/0/0 (ifindex 1) > virtio_net_hdr_sz 12 > features mask (0x): > features (0x58208000): >VIRTIO_NET_F_MRG_RXBUF (15) >VIRTIO_NET_F_GUEST_ANNOUNCE (21) >VIRTIO_F_ANY_LAYOUT (27) >VIRTIO_F_INDIRECT_DESC (28) >VHOST_USER_F_PROTOCOL_FEATURES (30) > protocol features (0x3) >VHOST_USER_PROTOCOL_F_MQ (0) >VHOST_USER_PROTOCOL_F_LOG_SHMFD (1) > > socket filename /var/run/vpp/sock1.sock type server errno "Success" > > rx placement: >thread 1 on vring 1, polling > tx placement: spin-lock >thread 0 on vring 0 >thread 1 on vring 0 >thread 2 on vring 0 > > Memory regions (total 2) > region fdguest_phys_addrmemory_sizeuserspace_addr > mmap_offsetmmap_addr > == = == == == > == == > 0 320x 0x000c 0x7f5a6a60 > 0x 0x7f47c420 > 1 330x0010 0x3ff0 0x7f5a6a70 > 0x0010 0x7f46f410 > > Virtqueue 0 (TX) > qsz 256 last_avail_idx 63392 last_used_idx 63392 > avail.flags 0 avail.idx 63530 used.flags 1 used.idx 63392 > kickfd 34 callfd 35 errfd -1 > > Virtqueue 1 (RX) > qsz 256 last_avail_idx 32414 last_used_idx 32414 > avail.flags 1 avail.idx 32414 used.flags 1 used.idx 32414 > kickfd 30 callfd 36 errfd -1 > > > > On Thu, Mar 22, 2018 at 3:07 PM, Sara Gittlin > wrote: > > i dont think these are error counters - anyway very poor pps > > > > On Thu, Mar 22, 2018 at 2:55 PM, Sara Gittlin > wrote: > > in the show err output i see that l2-output l2-learn l2-input counters > are continuously incremented : > show err >CountNode Reason > 11l2-output L2 output packets > 11l2-learnL2 learn packets > 11l2-inputL2 input packets > 3l2-floodL2 flood packets > > > * 8479644l2-output L2 output packets > 8479644l2-learnL2 learn packets8479644 >l2-inputL2 input packets* > > > > On Thu, Mar 22, 2018 at 11:59 AM, Sara Gittlin > wrote: > > Hello > i setup 2 vm connected to VPP as per the guide : > https://wiki.fd.io/view/VPP/Use_VPP_to_connect_VMs_Using_ > Vhost-User_Interface > > The performance looks very bad very low pps and large latencies > > udp pkt size 100B - throughput 500Mb > average latency is 900us > > i have 2 PMDs threads (200% cpu) in the host, in the VMs i see low > cpu load (10%) > > Please can you tell me what is wrong with my setup ? > > Thank you in advance > - Sara > > > > > > > > > > > >
Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser
Hi Sara, To address your question on “show err”, your output is displaying running counts for the l2-output, l2-learn and l2-input node wrt how many packets each node processed. These are not error counts, unless the “Reason” column indicate some kind of error. The CLI name “show err” is mis-leading in that it is showing stats kept by VPP nodes for various reasons, not just errors. I typically tell people to use the CLI “show node counters” which is more appropriate and produce the same output. The CLI “show err” was there for historical reasons and kept for backward compatibility. You are welcome to keep on using “show err” as it is shorter and faster to type. Regards, John From: vpp-dev@lists.fd.io On Behalf Of Sara Gittlin Sent: Thursday, March 22, 2018 8:56 AM To: vpp-dev@lists.fd.io Subject: Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser in the show err output i see that l2-output l2-learn l2-input counters are continuously incremented : show err CountNode Reason 11l2-output L2 output packets 11l2-learnL2 learn packets 11l2-inputL2 input packets 3l2-floodL2 flood packets 8479644l2-output L2 output packets 8479644l2-learnL2 learn packets 8479644l2-inputL2 input packets On Thu, Mar 22, 2018 at 11:59 AM, Sara Gittlin mailto:sara.gitt...@gmail.com>> wrote: Hello i setup 2 vm connected to VPP as per the guide : https://wiki.fd.io/view/VPP/Use_VPP_to_connect_VMs_Using_Vhost-User_Interface The performance looks very bad very low pps and large latencies udp pkt size 100B - throughput 500Mb average latency is 900us i have 2 PMDs threads (200% cpu) in the host, in the VMs i see low cpu load (10%) Please can you tell me what is wrong with my setup ? Thank you in advance - Sara
Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser
Hi Sara, Can you also send the results from show hardware and show interfaces? What are you using to test your performance. John From: on behalf of Sara Gittlin Date: Thursday, March 22, 2018 at 9:27 AM To: "vpp-dev@lists.fd.io" Subject: Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser this is the output of: show vhost-user VirtualEthernet0/0/0 Virtio vhost-user interfaces Global: coalesce frames 32 time 1e-3 number of rx virtqueues in interrupt mode: 0 Interface: VirtualEthernet0/0/0 (ifindex 1) virtio_net_hdr_sz 12 features mask (0x): features (0x58208000): VIRTIO_NET_F_MRG_RXBUF (15) VIRTIO_NET_F_GUEST_ANNOUNCE (21) VIRTIO_F_ANY_LAYOUT (27) VIRTIO_F_INDIRECT_DESC (28) VHOST_USER_F_PROTOCOL_FEATURES (30) protocol features (0x3) VHOST_USER_PROTOCOL_F_MQ (0) VHOST_USER_PROTOCOL_F_LOG_SHMFD (1) socket filename /var/run/vpp/sock1.sock type server errno "Success" rx placement: thread 1 on vring 1, polling tx placement: spin-lock thread 0 on vring 0 thread 1 on vring 0 thread 2 on vring 0 Memory regions (total 2) region fdguest_phys_addrmemory_sizeuserspace_addr mmap_offsetmmap_addr == = == == == == == 0 320x 0x000c 0x7f5a6a60 0x 0x7f47c420 1 330x0010 0x3ff0 0x7f5a6a70 0x0010 0x7f46f410 Virtqueue 0 (TX) qsz 256 last_avail_idx 63392 last_used_idx 63392 avail.flags 0 avail.idx 63530 used.flags 1 used.idx 63392 kickfd 34 callfd 35 errfd -1 Virtqueue 1 (RX) qsz 256 last_avail_idx 32414 last_used_idx 32414 avail.flags 1 avail.idx 32414 used.flags 1 used.idx 32414 kickfd 30 callfd 36 errfd -1 On Thu, Mar 22, 2018 at 3:07 PM, Sara Gittlin mailto:sara.gitt...@gmail.com>> wrote: i dont think these are error counters - anyway very poor pps On Thu, Mar 22, 2018 at 2:55 PM, Sara Gittlin mailto:sara.gitt...@gmail.com>> wrote: in the show err output i see that l2-output l2-learn l2-input counters are continuously incremented : show err CountNode Reason 11l2-output L2 output packets 11l2-learnL2 learn packets 11l2-inputL2 input packets 3l2-floodL2 flood packets 8479644l2-output L2 output packets 8479644l2-learnL2 learn packets 8479644l2-inputL2 input packets On Thu, Mar 22, 2018 at 11:59 AM, Sara Gittlin mailto:sara.gitt...@gmail.com>> wrote: Hello i setup 2 vm connected to VPP as per the guide : https://wiki.fd.io/view/VPP/Use_VPP_to_connect_VMs_Using_Vhost-User_Interface The performance looks very bad very low pps and large latencies udp pkt size 100B - throughput 500Mb average latency is 900us i have 2 PMDs threads (200% cpu) in the host, in the VMs i see low cpu load (10%) Please can you tell me what is wrong with my setup ? Thank you in advance - Sara
Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser
Sara, Could you please try again with adding below config to startup.conf? I remember some drivers don’t blast traffic at full speed unless it is getting a kick for each packet sent. I am not sure if you are using one of those. vhost-user { coalesce-frame 0 } Steven From: on behalf of Sara Gittlin Date: Thursday, March 22, 2018 at 6:27 AM To: "vpp-dev@lists.fd.io" Subject: Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser this is the output of: show vhost-user VirtualEthernet0/0/0 Virtio vhost-user interfaces Global: coalesce frames 32 time 1e-3 number of rx virtqueues in interrupt mode: 0 Interface: VirtualEthernet0/0/0 (ifindex 1) virtio_net_hdr_sz 12 features mask (0x): features (0x58208000): VIRTIO_NET_F_MRG_RXBUF (15) VIRTIO_NET_F_GUEST_ANNOUNCE (21) VIRTIO_F_ANY_LAYOUT (27) VIRTIO_F_INDIRECT_DESC (28) VHOST_USER_F_PROTOCOL_FEATURES (30) protocol features (0x3) VHOST_USER_PROTOCOL_F_MQ (0) VHOST_USER_PROTOCOL_F_LOG_SHMFD (1) socket filename /var/run/vpp/sock1.sock type server errno "Success" rx placement: thread 1 on vring 1, polling tx placement: spin-lock thread 0 on vring 0 thread 1 on vring 0 thread 2 on vring 0 Memory regions (total 2) region fdguest_phys_addrmemory_sizeuserspace_addr mmap_offsetmmap_addr == = == == == == == 0 320x 0x000c 0x7f5a6a60 0x 0x7f47c420 1 330x0010 0x3ff0 0x7f5a6a70 0x0010 0x7f46f410 Virtqueue 0 (TX) qsz 256 last_avail_idx 63392 last_used_idx 63392 avail.flags 0 avail.idx 63530 used.flags 1 used.idx 63392 kickfd 34 callfd 35 errfd -1 Virtqueue 1 (RX) qsz 256 last_avail_idx 32414 last_used_idx 32414 avail.flags 1 avail.idx 32414 used.flags 1 used.idx 32414 kickfd 30 callfd 36 errfd -1 On Thu, Mar 22, 2018 at 3:07 PM, Sara Gittlin mailto:sara.gitt...@gmail.com>> wrote: i dont think these are error counters - anyway very poor pps On Thu, Mar 22, 2018 at 2:55 PM, Sara Gittlin mailto:sara.gitt...@gmail.com>> wrote: in the show err output i see that l2-output l2-learn l2-input counters are continuously incremented : show err CountNode Reason 11l2-output L2 output packets 11l2-learnL2 learn packets 11l2-inputL2 input packets 3l2-floodL2 flood packets 8479644l2-output L2 output packets 8479644l2-learnL2 learn packets 8479644l2-inputL2 input packets On Thu, Mar 22, 2018 at 11:59 AM, Sara Gittlin mailto:sara.gitt...@gmail.com>> wrote: Hello i setup 2 vm connected to VPP as per the guide : https://wiki.fd.io/view/VPP/Use_VPP_to_connect_VMs_Using_Vhost-User_Interface The performance looks very bad very low pps and large latencies udp pkt size 100B - throughput 500Mb average latency is 900us i have 2 PMDs threads (200% cpu) in the host, in the VMs i see low cpu load (10%) Please can you tell me what is wrong with my setup ? Thank you in advance - Sara
Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser
this is the output of: show vhost-user VirtualEthernet0/0/0 Virtio vhost-user interfaces Global: coalesce frames 32 time 1e-3 number of rx virtqueues in interrupt mode: 0 Interface: VirtualEthernet0/0/0 (ifindex 1) virtio_net_hdr_sz 12 features mask (0x): features (0x58208000): VIRTIO_NET_F_MRG_RXBUF (15) VIRTIO_NET_F_GUEST_ANNOUNCE (21) VIRTIO_F_ANY_LAYOUT (27) VIRTIO_F_INDIRECT_DESC (28) VHOST_USER_F_PROTOCOL_FEATURES (30) protocol features (0x3) VHOST_USER_PROTOCOL_F_MQ (0) VHOST_USER_PROTOCOL_F_LOG_SHMFD (1) socket filename /var/run/vpp/sock1.sock type server errno "Success" rx placement: thread 1 on vring 1, polling tx placement: spin-lock thread 0 on vring 0 thread 1 on vring 0 thread 2 on vring 0 Memory regions (total 2) region fdguest_phys_addrmemory_sizeuserspace_addr mmap_offsetmmap_addr == = == == == == == 0 320x 0x000c 0x7f5a6a60 0x 0x7f47c420 1 330x0010 0x3ff0 0x7f5a6a70 0x0010 0x7f46f410 Virtqueue 0 (TX) qsz 256 last_avail_idx 63392 last_used_idx 63392 avail.flags 0 avail.idx 63530 used.flags 1 used.idx 63392 kickfd 34 callfd 35 errfd -1 Virtqueue 1 (RX) qsz 256 last_avail_idx 32414 last_used_idx 32414 avail.flags 1 avail.idx 32414 used.flags 1 used.idx 32414 kickfd 30 callfd 36 errfd -1 On Thu, Mar 22, 2018 at 3:07 PM, Sara Gittlin wrote: > i dont think these are error counters - anyway very poor pps > > On Thu, Mar 22, 2018 at 2:55 PM, Sara Gittlin > wrote: > >> in the show err output i see that l2-output l2-learn l2-input counters >> are continuously incremented : >> show err >>CountNode Reason >> 11l2-output L2 output packets >> 11l2-learnL2 learn packets >> 11l2-inputL2 input packets >> 3l2-floodL2 flood packets >> >> >> * 8479644l2-output L2 output packets >> 8479644l2-learnL2 learn packets 8479644 >>l2-inputL2 input packets* >> >> On Thu, Mar 22, 2018 at 11:59 AM, Sara Gittlin >> wrote: >> >>> Hello >>> i setup 2 vm connected to VPP as per the guide : >>> https://wiki.fd.io/view/VPP/Use_VPP_to_connect_VMs_Using_Vho >>> st-User_Interface >>> >>> The performance looks very bad very low pps and large latencies >>> >>> udp pkt size 100B - throughput 500Mb >>> average latency is 900us >>> >>> i have 2 PMDs threads (200% cpu) in the host, in the VMs i see low >>> cpu load (10%) >>> >>> Please can you tell me what is wrong with my setup ? >>> >>> Thank you in advance >>> - Sara >>> >>> >>> >>> >> > > >
Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser
i dont think these are error counters - anyway very poor pps On Thu, Mar 22, 2018 at 2:55 PM, Sara Gittlin wrote: > in the show err output i see that l2-output l2-learn l2-input counters > are continuously incremented : > show err >CountNode Reason > 11l2-output L2 output packets > 11l2-learnL2 learn packets > 11l2-inputL2 input packets > 3l2-floodL2 flood packets > > > * 8479644l2-output L2 output packets > 8479644l2-learnL2 learn packets 8479644 >l2-inputL2 input packets* > > On Thu, Mar 22, 2018 at 11:59 AM, Sara Gittlin > wrote: > >> Hello >> i setup 2 vm connected to VPP as per the guide : >> https://wiki.fd.io/view/VPP/Use_VPP_to_connect_VMs_Using_Vho >> st-User_Interface >> >> The performance looks very bad very low pps and large latencies >> >> udp pkt size 100B - throughput 500Mb >> average latency is 900us >> >> i have 2 PMDs threads (200% cpu) in the host, in the VMs i see low >> cpu load (10%) >> >> Please can you tell me what is wrong with my setup ? >> >> Thank you in advance >> - Sara >> >> >> >> >
Re: [vpp-dev] Very poor performance vm to vm via VPP vhostuser
in the show err output i see that l2-output l2-learn l2-input counters are continuously incremented : show err CountNode Reason 11l2-output L2 output packets 11l2-learnL2 learn packets 11l2-inputL2 input packets 3l2-floodL2 flood packets * 8479644l2-output L2 output packets 8479644l2-learnL2 learn packets 8479644 l2-inputL2 input packets* On Thu, Mar 22, 2018 at 11:59 AM, Sara Gittlin wrote: > Hello > i setup 2 vm connected to VPP as per the guide : > https://wiki.fd.io/view/VPP/Use_VPP_to_connect_VMs_Using_ > Vhost-User_Interface > > The performance looks very bad very low pps and large latencies > > udp pkt size 100B - throughput 500Mb > average latency is 900us > > i have 2 PMDs threads (200% cpu) in the host, in the VMs i see low > cpu load (10%) > > Please can you tell me what is wrong with my setup ? > > Thank you in advance > - Sara > > > >
[vpp-dev] Very poor performance vm to vm via VPP vhostuser
Hello i setup 2 vm connected to VPP as per the guide : https://wiki.fd.io/view/VPP/Use_VPP_to_connect_VMs_Using_Vhost-User_Interface The performance looks very bad very low pps and large latencies udp pkt size 100B - throughput 500Mb average latency is 900us i have 2 PMDs threads (200% cpu) in the host, in the VMs i see low cpu load (10%) Please can you tell me what is wrong with my setup ? Thank you in advance - Sara -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#8618): https://lists.fd.io/g/vpp-dev/message/8618 View All Messages In Topic (1): https://lists.fd.io/g/vpp-dev/topic/16048859 Mute This Topic: https://lists.fd.io/mt/16048859/21656 New Topic: https://lists.fd.io/g/vpp-dev/post Change Your Subscription: https://lists.fd.io/g/vpp-dev/editsub/21656 Group Home: https://lists.fd.io/g/vpp-dev Contact Group Owner: vpp-dev+ow...@lists.fd.io Terms of Service: https://lists.fd.io/static/tos Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub -=-=-=-=-=-=-=-=-=-=-=-