Speed tests results with AP

[Nas.CloudBusinessPortal]

So I have an RT-ax88U on 388.2.2, showing speed test over my expectations of 1.5gbe

Now an access point, RT-ax86u,on 388.2.2, teatherd to the rt-ax88u using the 2.5 connection.

I’m only getting 1000mbps speed test.

I’m a bit disappointed in the results.

Packet lost on the test says .5 to 1 percent loss

Any suggestion as to the speed difference.

I might go out a buy a second RT-AX88u as an access point , just to see if the difference is the RT-AX86u itself.

Doing a speed test from my iPhone XS Max, results in 500-550 mbps m, but it is an older iPhone.


Is there a router to router speed test?

 

[Nas.CloudBusinessPortal]

So

When I switched the Ethernet cable on the access point fron the 2.5 port to the 1.0 LAN port, the Speer tests decrease.

I get that.

What’s limiting the speed when the access point uses the 2.5 gbe port?

Is the RT-ax86u not good enough?

When I was using the RT-AX86u as the gateway router, it’s speeds too where not acceptable.

Comments?

ASUS claim this router to be a 2gbe capable.

 

[drinkingbird]

So

When I switched the Ethernet cable on the access point fron the 2.5 port to the 1.0 LAN port, the Speer tests decrease.

I get that.

What’s limiting the speed when the access point uses the 2.5 gbe port?

Is the RT-ax86u not good enough?

When I was using the RT-AX86u as the gateway router, it’s speeds too where not acceptable.

Comments?

ASUS claim this router to be a 2gbe capable.

Do not use the built in speed test, it is unreliable to say the least. Use a wired (preferable) or wireless PC. Obviously you'll be limited to the speed of that wired or wireless connection (around 60% of the connection speed when on wireless, and around 95% on wired).

 

[Nas.CloudBusinessPortal]

Do not use the built in speed test, it is unreliable to say the least. Use a wired (preferable) or wireless PC. Obviously you'll be limited to the speed of that wired or wireless connection (around 60% of the connection speed when on wireless, and around 95% on wired).

Lol my old laptop although connected at 846 mbps, only gets 100 mbps.

Oh well I’ll keep plugging at this issue.

I still think the ASUS RT-AX86u can’t go ant faster then 1100 mbps.

 

[drinkingbird]

Lol my old laptop although connected at 846 mbps, only gets 100 mbps.

Oh well I’ll keep plugging at this issue.

I still think the ASUS RT-AX86u can’t go ant faster then 1100 mbps.

Yeah "old" being the key word. Speed tests take a lot of CPU. Though if it is modern enough to have AC I'd think it could handle a lot more than that. Maybe too much stuff running on it, virus scanner interfering, etc. Or your wifi channel is heavily used.

I'm a bit confused, the AX88 does not have any 2.5 ports, only 1G. Are you using a 2.5 USB adapter off it? The AX86U only has one 2.5 port which would be for WAN or uplink, so no way to get >1G on a speed test on wired. Or are you using the AX88 pro?

Generally the AX86 comes highly recommended here, more so than the 88U (non-pro).

What happens if you use the AX86 as your main router with 2.5G for the WAN and run the built in speed test (not terribly reliable but may confirm whether it can get over 1G or not). You could also run a couple simultaneous tests from 1G wired devices and add the results.

 

[Nas.CloudBusinessPortal]

RT-AX88U Pro｜WiFi Routers｜ASUS Canada

ASUS gaming routers provide the very best gaming experience, with an arsenal of features and tools designed to improve online gaming performance and give you the competitive edge you need to win.

www.asus.com

Pro version has 2.5 wan and lan

When I used the RTAx86u as the main router, it could not generate the proper speed test.

So I’m still guessing that the Rt-Ax86u can’t do more that 1100mbps

Does anyone get speed test over 1.11gb ?

 

[drinkingbird]

RT-AX88U Pro｜WiFi Routers｜ASUS Canada

ASUS gaming routers provide the very best gaming experience, with an arsenal of features and tools designed to improve online gaming performance and give you the competitive edge you need to win.

www.asus.com

Pro version has 2.5 wan and lan

When I used the RTAx86u as the main router, it could not generate the proper speed test.

So I’m still guessing that the Rt-Ax86u can’t do more that 1100mbps

Does anyone get speed test over 1.11gb ?

The AX88U and AX88U Pro are totally different routers released years apart. That's why I asked if you were using the pro as it didn't make sense.

The 86U is similar to the 88U Pro and should be able to perform nearly as well, at least from other discussions I've seen here. Have you tried connecting it directly to the ISP and testing? Or use iperf and a couple machines on your network to test it.

 

[Nas.CloudBusinessPortal]

So today I bought a new rt-88u pro router

I set it up in AP mode.

It generated good results 1800-1900 mbps connected through two switches using a 2.5 gbe back to the main router, another rt-ax88u pro.

Packet loss was 1.5 percent, but that’s ok considering I’m using cat5e cable in my office/home.

So my conclusion…

My rt-ac86u as an access point, which only has a one gigabit land connection, gets 800mbps, I get that. Understandable.

But the rt-ax86u (not pro) only gets 1100-1200 mbps?

Quad core, 1.8 Ghz… hmmm

So how does ASUS able to advertise that this router has a peak data of 2 gbe?


All tests done with the radio off.

Any comments on the rt-ax86u using the 2.5 gb port not being able to transfer faster?

No wander I had issues with my isp getting slow speeds

 

[drinkingbird]

So today I bought a new rt-88u pro router

I set it up in AP mode.

It generated good results 1800-1900 mbps connected through two switches using a 2.5 gbe back to the main router, another rt-ax88u pro.

Packet loss was 1.5 percent, but that’s ok considering I’m using cat5e cable in my office/home.

So my conclusion…

My rt-ac86u as an access point, which only has a one gigabit land connection, gets 800mbps, I get that. Understandable.

But the rt-ax86u (not pro) only gets 1100-1200 mbps?

Quad core, 1.8 Ghz… hmmm

So how does ASUS able to advertise that this router has a peak data of 2 gbe?


All tests done with the radio off.

Any comments on the rt-ax86u using the 2.5 gb port not being able to transfer faster?

No wander I had issues with my isp getting slow speeds

Quad core 1.8ghz has very little to do with the speed. Routing uses a single core. If you are using full hardware acceleration the CPU should not really come into play. If you have features enabled that disable some or all acceleration, you're limited by the speed of 1 core basically and none of these routers will perform very well.

My router has a dual core 1ghz and with trend micro and other features enabled it caps out around 500M LAN to WAN. Your AX86 should probably get close to double that with similar features enabled, or with those features disabled probably close to 2 gig. If it is maxing at 1.1 to 1.2 with full hardware acceleration, there is something else you need to look at.

You should have 0 loss at 2.5 gig over Cat5e, so you need to look at your physical cabling and connectors. 1.5% packet loss can have a significant impact on throughput.

All that aside, advertising "up to 2 gig" is no different than saying "up to 100 meg". Both are "up to" and "in theory". As far as I know others have gotten much higher speeds on that model though.

 

[Nas.CloudBusinessPortal]

Quad core 1.8ghz has very little to do with the speed. Routing uses a single core. If you are using full hardware acceleration the CPU should not really come into play. If you have features enabled that disable some or all acceleration, you're limited by the speed of 1 core basically and none of these routers will perform very well.

My router has a dual core 1ghz and with trend micro and other features enabled it caps out around 500M LAN to WAN. Your AX86 should probably get close to double that with similar features enabled, or with those features disabled probably close to 2 gig. If it is maxing at 1.1 to 1.2 with full hardware acceleration, there is something else you need to look at.

You should have 0 loss at 2.5 gig over Cat5e, so you need to look at your physical cabling and connectors. 1.5% packet loss can have a significant impact on throughput.

All that aside, advertising "up to 2 gig" is no different than saying "up to 100 meg". Both are "up to" and "in theory". As far as I know others have gotten much higher speeds on that model though.

It’s all AP for this testing, with wireless turned off, no features installed.

 

[drinkingbird]

It’s all AP for this testing, with wireless turned off, no features installed.

Are you using the internal speed test of the router? If so - that's most likely your issue. That uses 1 core of the CPU and generating that type of traffic needs a lot of power. If you're in AP mode, the Asus is just acting as a switch, but with only one 2.5G port you're not going to be able to max it out with a single device, you'd want to run 2 simultaneous tests from dual 1G connected PCs. If you can't hit about 1.8 to 1.9 gig that way, time to start checking your cabling and the PCs you're using for testing.

 

[Nas.CloudBusinessPortal]

Are you using the internal speed test of the router? If so - that's most likely your issue. That uses 1 core of the CPU and generating that type of traffic needs a lot of power. If you're in AP mode, the Asus is just acting as a switch, but with only one 2.5G port you're not going to be able to max it out with a single device, you'd want to run 2 simultaneous tests from dual 1G connected PCs. If you can't hit about 1.8 to 1.9 gig that way, time to start checking your cabling and the PCs you're using for testing.

Test are identical.

So one model can’t generate 2gbe and the the others can.

It pretty simple.

Where do you get your information from about Speedtest using on core only?

Your not getting the picture.

 

[drinkingbird]

Test are identical.

So one model can’t generate 2gbe and the the others can.

It pretty simple.

Where do you get your information from about Speedtest using on core only?

Your not getting the picture.

Nowhere are the routers advertised as being able to generate 2 gig of speedtest traffic. It is well known all through these forums that the internal speed test is unreliable. It relies completely on the very low powered CPU and will hit a limitation at higher speeds. A quad 1.8ghz ARM is not a powerful CPU.

Watch the CPU while the speed test is running, you'll see it locked to a single core. The linux base running on these routers doesn't support a single task using multiple threads.

 

[Nas.CloudBusinessPortal]

Watch the CPU while the speed test is running, you'll see it locked to a single core. The linux base running on these routers doesn't support a single task using multiple threads.

According to your theory, using a single core,

Then my rt-ax88u pro can generate 1.9 gbe (or more, cause my isp limits me ) of internet traffic

And my rt-a 86u can generate only 1.2 maximum of internet traffic.

My question is why?

 

[drinkingbird]

According to your theory, using a single core,

Then my rt-ax88u pro can generate 1.9 gbe (or more, cause my isp limits me ) of internet traffic

And my rt-a 86u can generate only 1.2 maximum of internet traffic.

My question is why?

RT-AX86U - quad-core 1.8GHz Broadcom BCM4908 SoC with ARM Cortex A53
88U Pro - BCM4912 SOC with ARM B53 quad core 2.0Ghz. BCM4912 lists improved packet processor and various other improvements (it isn't just clock speed that matters, especially on these ARM CPUs).

These routers are designed primarily to forward/pass traffic, not run high load applications on the system such as generating/processing a lot of dummy data for a speed test. Speed test for each model will be somewhat reliable up to a certain speed (differs by model). But even when the CPU isn't a bottleneck, people have found issues with the test not being accurate.

 

[Nas.CloudBusinessPortal]

So, Ethernet testing using an

Acer switch (SA)-271 laptop , intel i5-6200U at 2.3 ghz

Using the usb 3.0 dongle to 2.5 Ethernet.

I get 1.3 Gbps.

So this test tells me…

Using a 2.5 gbe dongle is limited?


Did a massive copy from the NAS drive 29.2 GB and peeked out at 273 MB/s or 2.111 Gbps , so that sounds right.

 

[drinkingbird]

So, Ethernet testing using an

Acer switch (SA)-271 laptop , intel i5-6200U at 2.3 ghz

Using the usb 3.0 dongle to 2.5 Ethernet.

I get 1.3 Gbps.

So this test tells me…

Using a 2.5 gbe dongle is limited?


Did a massive copy from the NAS drive 29.2 GB and peeked out at 273 MB/s or 2.111 Gbps , so that sounds right.

I'm assuming the NAS was LAN to LAN using the USB dongle? If so you know that dongle can do over 2. Are you still using that cable that was seeing 1.5% packet loss for your WAN speed tests?

If you factory reset the 86 with minimal config and plug 2.5G into ISP and 2.5G into a PC, what are the results? Use short, known good patch cables. Watch the CPU on PC and router and make sure nobe of the cores hit 100%. On the PC you have to change task manager to show all cores separately not the average.

All else fails try an older firmware, maybe something has gotten messed up in the newer ones.

 

[Nas.CloudBusinessPortal]

RT-AX86U - quad-core 1.8GHz Broadcom BCM4908 SoC with ARM Cortex A53
88U Pro - BCM4912 SOC with ARM B53 quad core 2.0Ghz. BCM4912 lists improved packet processor and various other improvements (it isn't just clock speed that matters, especially on these ARM CPUs).

Speed test for each model will be somewhat reliable up to a certain speed (differs by model).

That’s what I’m thinking. Speedtest are reliable, as long as the router can generate the traffic.

As you mentioned, only one core generate tragic on sped tests; but, have you also noticed that a router’s cpu cores only use one core for passing data from the lan to wan?

My claim still stands.

If the router can generate internet traffic fast enough for a speed test for 1gbps then it can’t handle 1 Gbps from wan to lan as a switch either.

I’m challenging ASUS on this, and asking to prove me wrong.

 

[Nas.CloudBusinessPortal]

What’s your router model and what’s the fastest lan to wan transfer you have witnessed?

 

[drinkingbird]

That’s what I’m thinking. Speedtest are reliable, as long as the router can generate the traffic.

As you mentioned, only one core generate tragic on sped tests; but, have you also noticed that a router’s cpu cores only use one core for passing data from the lan to wan?

My claim still stands.

If the router can generate internet traffic fast enough for a speed test for 1gbps then it can’t handle 1 Gbps from wan to lan as a switch either.

I’m challenging ASUS on this, and asking to prove me wrong.

Except passing traffic/routing/NAT is accelerated by a hardware ASIC assuming you don't have features enabled that disables that. Generating traffic (and most other software functions like aiprotection, parental controls, QOS, VPN, etc) are either partially accelerated or not at all.

Routing typically will use a different CPU from applications too.

My AC1900 can pass close to 1Gbps with all those features disabled, and around 500 with aiprotection etc enabled. In the second scenario, NAT acceleration is enabled but hardware/routing is not.