I think as you get above 1Gbps everything is very expensive and the computational costs are very high. It can make more sense to use simpler queuing systems with very high bandwidth.
For example I have my desktop computers using QFQ for their gigabit NICs. Then my switches use weighted round robin, then my router uses HFSC.
But I have only gigabit internet. What would I do with 2.5Gbps or 5Gbps internet? I would need faster switches, and 10Gbps switches are ~ $500 by themselves. It doesn't make sense, the value compared to gigabit is too limited. Most of the time websites can't even saturate my gigabit. Speed test sites struggle to hit gigabit even.
So I do think it's best to not bother with speed tiers above gigabit, at least until 10GbE switches are ~$150 or so
I have done something like that, and it did have benefits.
Back in the day, when I ran my TP-Link C7 as my router, I upgraded to a faster link, and I found out that a C7 didn't have the horsepower to SQM my new 300/30mbit cable link. It ran out of resources at about 115mbit, with wifi, actually.
Not wanting to lose that much of my new download speed, for quite a while I just SQM'ed the upload side, which being 30mbit, was no big load. I noticed that it did quite a lot to decrease the download latency as well, and helped overall a lot more than I expected. So, I'd guess just shaping your 600mbit up, might also give you some improvement... on something you might have today that can shape at that rate.