802.11ax Routers

For WiFi 6 80MHz (1024QAM) each stream should support 600.5Mbps.

For WiFi 5 80MHz (256QAM) each stream should support 433.3Mbps.

Between 2x wirelessly connected RT3200's on separate floors in my house:

root@OpenWrt:~# iperf3 -c 192.168.1.2
Connecting to host 192.168.1.2, port 5201
[  5] local 192.168.1.1 port 45890 connected to 192.168.1.2 port 5201
[ ID] Interval           Transfer     Bitrate         Retr  Cwnd
[  5]   0.00-1.00   sec  91.2 MBytes   765 Mbits/sec    0   1.95 MBytes
[  5]   1.00-2.00   sec  93.4 MBytes   781 Mbits/sec    0   1.95 MBytes
[  5]   2.00-3.00   sec  93.8 MBytes   787 Mbits/sec    0   1.95 MBytes
[  5]   3.00-4.00   sec  89.8 MBytes   754 Mbits/sec    0   1.95 MBytes
[  5]   4.00-5.00   sec  88.6 MBytes   743 Mbits/sec    0   1.95 MBytes
[  5]   5.00-6.00   sec  91.0 MBytes   761 Mbits/sec    0   1.95 MBytes
[  5]   6.00-7.02   sec  91.1 MBytes   755 Mbits/sec    0   1.95 MBytes
[  5]   7.02-8.00   sec  85.0 MBytes   724 Mbits/sec    0   1.95 MBytes
[  5]   8.00-9.00   sec  93.6 MBytes   785 Mbits/sec    0   1.95 MBytes
[  5]   9.00-10.00  sec  86.0 MBytes   721 Mbits/sec    0   1.95 MBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bitrate         Retr
[  5]   0.00-10.00  sec   904 MBytes   758 Mbits/sec    0             sender
[  5]   0.00-10.01  sec   902 MBytes   756 Mbits/sec                  receiver

iperf Done.
2 Likes

That is a good result, how many 80MHz streams are these devices using?

I think the RT3200 can support up to 4 streams.

Yes, the RT3200 is 4T4R. Be aware that only the 5 GHz radio (MT7915) does WiFi 6, but the 2.4 GHz is still WiFi 4 only (ie. 802.11bgn), but also 4T4R (it probably could do out-of-spec WiFi 5 VHT rates on 2.4 GHz, but almost no clients support that).

Even if the hardware supports, it must be a proprietary implementation as VHT on 2.4GHz is not in compliance with the IEEE 802.11 standard.

That is why few to none clients will support it.

https://www.cwnp.com/forums/posts?postNum=307375

There is an old pull request for it. I can see this mostly benefiting Mesh networks with 2.4GHz backhaul.

2 Likes

Mods for ZBTLINK Z100-a with 32M SPI Flash

You can build after with samba4

would it be possible to get openwrt going on the BCM6755 sans wifi
really don't care about the wifi capability at all
the stock firmware of the archer AX1800 V1.20 is based on Openwrt 12 Attitude adjustment
I have it extracted and dumped if anybody wants it shoot me a pm

That chip is so new I doubt you will find much mainline Linux kernel support for it as of yet.

…and given the wireless situation, there probably won't be many who'd buy (let alone work on porting for-) this device either.

1 Like

blarg there is nothing affordable in the consumer hardware class that isn't A: underpowered or B: running mips (yuk)

Are you programming in assembly? Developing compiler backends? Maintaining a separate CI infrastructure just for OpenWRT on MIPS?

If not, why do you care what the ISA is as long as it offers the performance you need?

thats the thing mips is horrendously slow
recursive dns perf on mips sucks

The mt7622 platform is currently one of the fastests platforms available. It can even shape SQM at linerate (1 Gbit/s). Old MIPS models may be slow, but this one definitely is not. Just focus on actual benchmarks instead of the architecture. If it performs well enough for your requirements, then just use it instead of dismissing it merely due to its architecture.

Edit: disregard, mt7622 is actually ARM and not MIPS. Both use a family of the RISC instruction set.

1 Like

Not with CAKE though right?

No, only fq_codel

2 Likes

It looks like it's part of Networking Pro 1610 Platform

Now that is a beast, IPQ8078A with two QCN9074-s, and finally, somebody put a SFP+ slot on one of the 10G ports(Even heatsinked for GPON-s).

I am afraid to even see what the price is

2 Likes

Supposedly there's an even better model coming
d43c-kmkptxc4470052

outrageously expensive about $900 !

Really interesting specs, but the thermal- and rf design probably won't win (m)any prices, somehow I don't think that's a good approach to it.