10GBE Peer to Peer Connection (TrueNAS Scale - Windows10)

[DigitalMinimalist]

Now I have to optimize the connection - it currently maxes out at 500-600MB/s - even with NVME to NVME

I already followed a guide to optimize buffers and checked the PCIe lanes:
X710 has a PCIe x8 interface with max 8Gb/s - one card in x4 slot the other in x8, which should both not bottleneck at 500 MB/s

Any other typical optimization possible?