Tuning OpenWRT routing performance

Hi, I've tried to measure the performance of the Raspberry Pi 4B, using iperf and iperf3. The TCP throughput is in the range of 804-810 Mbps, while UDP throughput can hit 1 Gbps. I use two different hosts that I send iperf3 and iperf tests to: one is a 2012 Mac Mini (Quad Core i7) and the other one is a QNAP NAS (Quad Core AMD). As I start increasing the number of iperf streams, the throughput to the QNAP starts dropping with every additional stream, but the throughput to Mac Mini stays about the same (or improves marginally by a couple Mbps).

These numbers are without involving NAT. Also, I placed both interfaces in the same firewall zone. I've enabled software and hardware offloading for routing. I've also selected the "Enable packet steering across all CPUs. May help or hinder network speed" check box. None of those settings changed the throughput numbers.

Incidentally, when I run iperf3 directly between my MacBook Pro and the Mac Mini, I get 941 Mbps. The iperf3 test from the MacBook Pro to the QNAP is 940 Mbps. So, I know that the Raspberry Pi 4B cuts the throughput from 940 Mbps to 806 Mbps.

So, I would like to get my TCP throughput through the OpenWRT on the Raspberry Pi4B to 940 Mbps. Can anyone advise what could be done to improve the throughput or point me in the right direction?

Thank you.