Hello Vivek, There was Kubernetes/Calico-VPP performance benchmark done with / without DMA support for memif ( https://www.tigera.io/blog/high-throughput-kubernetes-cluster-networking-with-the-calico-vpp-dataplane-and-accelerated-memif/). Since Calico-VPP was also tested under Marvell, it may be good use case for your benchmark. Jerome
De : [email protected] <[email protected]> de la part de kvivekk via lists.fd.io <[email protected]> Date : vendredi, 28 novembre 2025 à 07:59 À : [email protected] <[email protected]> Objet : [vpp-dev] Measuring CPU Cycles in VPP Hello everyone, I’m currently evaluating a set of test cases to benchmark session-layer acceleration—with and without DMA. Beyond measuring standard throughput using scenarios such as echo server–client or static server tests, I’m looking for a robust test setup that will allow me to specifically measure CPU cycles spent on data copying in either case. If you have suggestions or experience with concrete methods or test cases that highlight copy-related CPU overhead, I’d greatly appreciate your input. Thank you!! Regards, Vivek
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#26576): https://lists.fd.io/g/vpp-dev/message/26576 Mute This Topic: https://lists.fd.io/mt/116510533/21656 Group Owner: [email protected] Unsubscribe: https://lists.fd.io/g/vpp-dev/leave/14379924/21656/631435203/xyzzy [[email protected]] -=-=-=-=-=-=-=-=-=-=-=-
