SQM makes bufferbloat worse

[accelerate]

You could try to use a build with fast path included.

Qualcomm Fast Path For LEDE

Ok On WDR4300 OC to 730MHZ Up Down limit on eth0.2 interface set to 800000kbit/s = 800Mbits jperf cake layer_cake = 571 Mbits/sec jperf fq_codel simple.qos = 663 Mbits/sec Throughput definitely increased but not line rate. I guess lower cpu utilization allows better SQM So to answer the question: No you will not get line rate with SQM even with fast-path but performance will be better than without fast-path How much performance you get with SQM still depends on the CPU.

Qualcomm Fast Path For LEDE

I maxed out ingress and egress at 800Mbits Since Fast-Path can do 900Mbit cake+layer_cake = 571 Mbits/sec fq_codel+simple.qos = 663 Mbits/sec I am using GUI to test on jperf On dslreport speedtest I can get A+ for google chrome but not for firefox. I think the service is quite cranky and inconsistent.

It seems like it also does increase SQM performance.
In case you want to try it:

GitHub - gwlim/openwrt-sfe-flowoffload-ath79: Openwrt firmware with SFE and...