How Are You Optimizing Packet Processing Performance on QorIQ T-Series Platforms?

[jeremiranner]

Has anyone worked on optimizing packet processing performance on QorIQ T-Series platforms under heavier network traffic conditions?

We are currently testing a T-Series based system for continuous packet forwarding and noticed that CPU utilization increases quite quickly once traffic bursts become sustained. The platform remains stable, but latency consistency between cores starts varying under load.

I am particularly interested in understanding how others are handling:

interrupt distribution
queue balancing
multicore workload tuning
memory/cache optimization
latency reduction during high throughput

Would appreciate hearing about any practical tuning methods or performance observations from real deployments.

[Bio_TICFSL]

Hello,

Yes — there is documented guidance for exactly the kinds of tuning you listed on QorIQ T-Series/DPAA1 systems, in the LA1224-RDB-BSPUG and it aligns closely with the behavior you are seeing under sustained bursts: as load rises, the critical tradeoff becomes throughput vs. per-core latency consistency . The retrieved documentation points to five main levers: interrupt moderation, FQ/channel distribution policy, dequeue/QMI tuning, cache stashing/CPC usage, and order-handling mode selection.

Regards