I haven't found much information but I believe it has a 320Mhz CPU, my connection is not top tier either: 25 Mbps download and 2 Mbps upload.
I have succesfully patched this firmware: https://openwrt.org/toh/adb/p.dg_a4001n1 and I'm running SQM Cake with these settings: https://openwrt.org/docs/guide-user/network/traffic-shaping/sqm
I'm just unsure if this might be too much for the poor thing as sometimes when playing Apex Legends I get some random lag spikes.
If this is the case then. Which good router do you recommend for SQM cake and my current internet speed?. Which settings should I use on SQM or should I disable it completely?
A typical test for overload would be to generate saturating traffic, say a number of wget down and uploads, and/or uploads to mozilla's encrypted file sharing service (https://send.firefox.com while logging into the router via secure shell (ssh) and monitor the idle and sirq fields of a concurrently started
top -d 1. If idle gets too close to zero (rule of thumb if you see <= 5% idle things are probably too tight for comfort.
That way you will not need to extrapolate from the experience others have with their routers...
Well I did some speedtests and it got close to 0% idle for a split second, mostly when downloading, after that it stayed at 15-30% idle. I think the cause of my lag spikes while playing is just poor infrastructure from my ISP, I do get lag spikes even when pinging www.google.com and being 95% idle with WiFi disabled.
Ah, in case of expected latency spikes from the ISP's side I would run mtr on the router (
opkg update ; opkg install mtr ; mtr -z 220.127.116.11) and look how the worst case RTTs are distributed along the path. You can/should replace 18.104.22.168 (Goggle's DNS server, which will typically resolve to a close by data center) with something more relevant for your usage. If you find something suspicious there it would be time to try to also get reverse traceroutes/mtrs from the target site to your router going. Please be aware that interpretation of mtr/traceroure results is less straight forward than one would hope (see https://archive.nanog.org/sites/default/files/10_Roisman_Traceroute.pdf for some pointers).