Thanks for your thoughts. The comment about SDK got me thinking, and now I suspect the issue is MikroTik’s ability, reluctance or lack of interest to reimplement the NSS accelerator drivers required for the ipq-8064 to achieve 1Gbps under their ros7/kernel 5.6 build.
Oh well, it seems you are right; this will never happen and we are stuck with RouterOS 6 on this model which is still being sold today.
All this being said an official note about really bad performance under 7 would be nice, instead of dead silence and no word what so ever from the manufacturer.
The NSS driver is probably closed source, just like Broadcom’s acceleration bits. Again, those drivers are surely for a specific kernel version and must be updated by the SoC manufacturer. So basically, MT can’t update or use that driver - they rely on FastTrack for NAT and L3 HW offloading on the PHY (limited to CRS/CCR models) to decrease CPU load. To be fair, that SoC was announced in 2013 - selling new parts makes a lot more sense when transistor density increases and production cost/transistor count decreases over time. It is planned obsolescence…
MikroTik did reply to complaints about poor performance under ROS 7 with a detailed explanation in the past:
The thread you linked includes this reply from MikroTik:
We can also calculate the cache-hit ratio (x) for “512-byte / 25 ip filter rules beyond” case where v6 routing could perform faster than v7:
(100 - x) + 1.6x = 120
0.6x = 20
x = 33
Hence, v6 and v7 routing performance match if 33% of the packets hit the routing cache on v6. This is clearly unrealistic, unless there is a small number of connections going through a high-end device (e.g., using CCR1036 in SOHO network).
I argue that the v7 numbers for the RB3011 do not reflect the above statement and the issue on the RB3011 in particular goes beyond a simple “the routing cache is disabled” explanation.
I am resigned that MikroTik is unable to make the RB3011 perform in v7. The 880mhz hex is basically equivalent to it (or better?) on v7.
Anyway, I give up at this point as they are obviously not going to fix it. If the hardware as you say is considered obsolete then they should confirm that and move it to EOL status.