Does differing host/network byte ordering on AGX add to network latency?

[jaiyamsharma]

I am trying to make two Jetson AGXs communicate via ethernet with as low latency as possible using UDP protocol. The default request-response latency measured by netperf is around 200 microseconds. I am looking for ways to reduce this and all suggestions are welcome.

Looking at the network stack, I came across the fact that Jetson uses little endian byte order while network uses big endian order. So, for a request response scenario, byte order conversion needs to be done 4 times

–>start
Host sender (LE)–> BE → send to client
Client receiver BE → LE
Client sender LE–> BE -->send to host
Host receiver BE → LE
→ end

Of course this is a very simplified picture and I have omitted all parts of the stack unrelated to byte ordering. My question is, does this 4 time conversion impact the latency significantly? If one were to use systems with big endian ordering, given everything else remains same, would that reduce network latency by any measurable amount?

[WayneWWW]

Please check if adding CONFIG_EQOS_DISABLE_EEE could help.

I don’t think LE/BE could be an issue here.

[jaiyamsharma]

@WayneWWW Thanks. Where do I add this line?

[WayneWWW]

This is inside the kernel. There is a tegre_defconfig file to define which options are enabled/disabled in our kernel driver.

Download the kernel source code from

https://developer.nvidia.com/embedded/linux-tegra-archive

Follow the document here to build the kernel on the host.

https://docs.nvidia.com/jetson/l4t/index.html#page/Tegra%20Linux%20Driver%20Package%20Development%20Guide/kernel_custom.html#wwpID0E0WD0HA

Replace your kernel to /boot/Image on the device and reboot.

And just some precaution that rebuilding kernel has chance to crash the system so using uart serial console might be helpful.

NVIDIA Xavier - Serial Console