Question AQtion 10Gbit Network Adapter. Speed is only 2GB ?

[Victel]

I have the Z590 AORUS XTREME WATERFORCE with the 10gb network adapter so it should be capable of 10gb. I had ATT come out and check that the slow speeds weren't being caused by them. They said it's my hardware limiting me. I updated my network adapter to the latest version from their official website: https://www.marvell.com/support/downloads.html# still only getting 2gb of speed. I have the fiber plan and I'm using cat 8 cables. Any ideas? Thanks

 

[USAFRet]

Moved to Networking

 

[Grobe]

I have the Z590 AORUS XTREME WATERFORCE with the 10gb network adapter so it should be capable of 10gb.

And what other equipment? (can the issue be with the network adapter in the other end of the cable? )

still only getting 2gb of speed.

How did you test this (detailed)? What (if any) web service used? Tried another (web server or service)? Same result or different?

I have the fiber plan

Ok, can you share (any personal info removed) (assuming this is some sort of printable drawing you talking about) ?

Reason I ask is because I suspect the way you test may be the reason of lower transmission speed.

 

[Victel]

And what other equipment? (can the issue be with the network adapter in the other end of the cable? )

How did you test this (detailed)? What (if any) web service used? Tried another (web server or service)? Same result or different?

Ok, can you share (any personal info removed) (assuming this is some sort of printable drawing you talking about) ?

Reason I ask is because I suspect the way you test may be the reason of lower transmission speed.

I apologize, i forgot to mention my router is the BGW320, the model AT&T provided for my speed. Is that the question you're asking?

I tested it on speedtest.net and the AT&T U-verse web speed tester. AT&T came out and tested it on their end and the signal going to my PC and they said it's fine.

 

[bill001g]

It is highly likely the cat8 cable you have is a fake. Real cat8 cables are extremely expensive and only used in large data centers where they have 40gbit ports. I have not ever seen real ones sold on places like amazon you tend only get them from specialty places.

It might be ok though,depends if they used thin wires. You only really need cat5e to run 2.5 and 5g. Normal cat6 will do 10gbit at shorter distances. The cable you really want is cat6a BUT you are not running 10gbit anyway.

If you look at the status of the port is it connecting at 2.5g or 5g. What plan do you pay att for. If you go into the router and run the speedtest from the testing panel in the router itself what speed do you get.

 

[Victel]

The download speed is above 5500 on the router speed test. Says port one transmit speed is 500000000 port 2,3,4, are all 100000000. Is that what you're asking about? I pay for the 5gb plan, $180 a month

 

[bill001g]

This tends to be very hard to test since you need multiple machine and ports that can run 5gbit.

Since the ATT router itself gets the full speed and your machine is connecting at 5gbit that eliminates the easy things to check. Normally I would say test with a tool like IPERF but you need a second machine in your house that can run 5gbit and a switch/router that has addition 5gbit ports.
Maybe watch the resource manager while the speedtest is running and see if maybe you are maxing out something. You wouldn't think it takes much cpu etc to transfer a file that it does not even save.

I guess you could try to boot a linux USB test image. Most have a browser preloaded in the default install and most have drivers for the nic card you are using. This though only help a little since it would just confirm you do not have defective hardware it does not tell you what silly setting in windows might be causing the problem.

I have att and every now and then they bother me about upgrading to 2.5 or 5g. If there was more cost difference I might consider going slower than the 1gbit plan. When it was first installed they had it at 2.5g for a week or so. I found most site...like steam...did not really even want to use 1gbit speeds. It is hard to find a site that lets you download at really fast speeds. I really don't download that much on a daily or even weekly basis to justify $180 month 5gbit plan.

 

[Victel]

This tends to be very hard to test since you need multiple machine and ports that can run 5gbit.

Since the ATT router itself gets the full speed and your machine is connecting at 5gbit that eliminates the easy things to check. Normally I would say test with a tool like IPERF but you need a second machine in your house that can run 5gbit and a switch/router that has addition 5gbit ports.
Maybe watch the resource manager while the speedtest is running and see if maybe you are maxing out something. You wouldn't think it takes much cpu etc to transfer a file that it does not even save.

I guess you could try to boot a linux USB test image. Most have a browser preloaded in the default install and most have drivers for the nic card you are using. This though only help a little since it would just confirm you do not have defective hardware it does not tell you what silly setting in windows might be causing the problem.

I have att and every now and then they bother me about upgrading to 2.5 or 5g. If there was more cost difference I might consider going slower than the 1gbit plan. When it was first installed they had it at 2.5g for a week or so. I found most site...like steam...did not really even want to use 1gbit speeds. It is hard to find a site that lets you download at really fast speeds. I really don't download that much on a daily or even weekly basis to justify $180 month 5gbit plan.

I do currently max out my download speed when I download stuff, with 5gb I can download many files at once. What settings in my PC actually affect how it receives my internet?

 

[bniknafs9]

maybe the ports on your router are limiting you ?

 

[Victel]

maybe the ports on your router are limiting you ?

How do I check that and fix it?

 

[bill001g]

It is not likely going to be some specific setting limiting you. There is software that limit data rates so that say a background download does use 100% of your bandwidth. You generally know when you have set that up and I doubt it would limit it to 2gbit, it is more when you only have a 50mbps internet that you can easily exceed.
A example would be the download limit you can set in steam.

There are all kinds of things that can limit your speeds that only show up when you are running extremely fast internet connections. For example the tcp window size can limit the maximum rates you can transfer. This is partially why speedtest uses multiple parallel data download streams. It can be all kinds of stuff like some of the settings in the nic that offload certain features from the cpu to the nic. Sometime they are better on and other times better off.

Your largest issue is you are testing to the internet where you have no control of the remote end. You never really know for sure if the server is limiting you say because it is busy or if it is purely your machine doing the limit. Since you do not have the ability to test in your house between machine you have 100% control over it is goind to be tricky to get good test data. The loads on servers on the internet can change between your test runs.

Even if you get it to work what I suspect you will find is sites like speedtest will show you big numbers but other sites you actually download from will have artificial limits on your. The server owners only have so much bandwidth and they place caps on the rates so a small number of users with very fast internet connection do not block all the other customers. I know steam only very rarely lets be download anywhere close to my 1gbit internet speed.

 

[Victel]

It is not likely going to be some specific setting limiting you. There is software that limit data rates so that say a background download does use 100% of your bandwidth. You generally know when you have set that up and I doubt it would limit it to 2gbit, it is more when you only have a 50mbps internet that you can easily exceed.
A example would be the download limit you can set in steam.

There are all kinds of things that can limit your speeds that only show up when you are running extremely fast internet connections. For example the tcp window size can limit the maximum rates you can transfer. This is partially why speedtest uses multiple parallel data download streams. It can be all kinds of stuff like some of the settings in the nic that offload certain features from the cpu to the nic. Sometime they are better on and other times better off.

Your largest issue is you are testing to the internet where you have no control of the remote end. You never really know for sure if the server is limiting you say because it is busy or if it is purely your machine doing the limit. Since you do not have the ability to test in your house between machine you have 100% control over it is goind to be tricky to get good test data. The loads on servers on the internet can change between your test runs.

Even if you get it to work what I suspect you will find is sites like speedtest will show you big numbers but other sites you actually download from will have artificial limits on your. The server owners only have so much bandwidth and they place caps on the rates so a small number of users with very fast internet connection do not block all the other customers. I know steam only very rarely lets be download anywhere close to my 1gbit internet speed.

So what do I do in order to figure this out? Is there a more reliable speed test?

 

[bill001g]

There are other speedtesting sites but many are not as good as ookla. I have seen some that only get say 200mbps.

It is not so much reliable it is you need to find a way to eliminate variables. You know that the speedtest site itself can do the speeds because the router can test to it. Be sure you test to the same exact speedtest server. ATT uses one that is on their own network when it tests from the router.
When you run speedtest for example you are using a browser that could have any number of different settings that could affect it. That is before you even think to look at windows settings.

So IF you had another machine the normal tests I would recommend was first to run IPERF between 2 machines in your house. This is a simple line mode tool. It pretty much tests the drivers and nic only. You would then use file sharing and attempt to copy very large files between machines in your house. This is more complex and uses more of the OS but is not as complex as internet browser based download. If all this is good on the LAN/LAN side you would then suspect maybe a router setting. When the router does speedtest it runs directly on the public IP it does not go through the NAT overhead.

You are almost going to have to blindly try changing stuff you have no way to actually test without having everything on the internet and your ISP in your testing path. You can be fairly certain it is a setting in the pc because of the good results on the router

 

[bniknafs9]

How do I check that and fix it?

Check it with your router's manufacturers , check my other thread too . i bought a new ONT together with a 2.5 Gbps network interface . but it only gives me 1 Gbps . i think using the latest 5 Ghz wifi also wifi 6 and 7 gives you faster speed s.

 

[Victel]

Check it with your router's manufacturers , check my other thread too . i bought a new ONT together with a 2.5 Gbps network interface . but it only gives me 1 Gbps . i think using the latest 5 Ghz wifi also wifi 6 and 7 gives you faster speed s.

I will try with them and I also have an open ticket with my motherboard, maybe they can help.

 

[Grobe]

I will try with them and I also have an open ticket with my motherboard, maybe they can help.

Please remember to reply back when you got a response (confirms bug, or something else).

 

[Victel]

Please remember to reply back when you got a response (confirms bug, or something else).

I'm still waiting on them. They're incredibly slow and don't seem to be asking the correct questions and they only reply once per day and not on the weekends. I spoke with them on the phone and they were busy asking me about the intel network adapter on my PC (my pc supports two and the AQC one supports 10g while the one they were asking me about only supports 2.5)

In my other findings I did discover according to my speed tests that my upload is signifanctly faster than my download at 3.5g where my download is 2g. I also checked in my network settings on windows and it shows 5g, same with checking it in command prompt running:
wmic nic where netEnabled=true get name, speed

Maybe these are more accurate than the web speed tests such as speedtest.net?

 

[bill001g]

All those commands mean is that the port electrically is running at 5gbit. Very technically it is always transferring at 5gbit. A ethernet port either transmits at full speed or zero. Those numbers are not real useful so most software measure data transfer over a period of time and take a average. The problem is there is something that is capping the throughput at a lower rate than the physical cable can do. A example would be say you run a vpn on your router. It might have gigabit wan ports and the ISP plan might be gigabit but the overhead of the encryption will put so much load on the cpu that you are lucky to get 30-50mbps on most router. Hard to say in the case of a pc since it is unlikely a cpu bottleneck. It almost has to be some setting or driver issue that is causing it to limit less than what the physical media can do.

Again this is why it would be nice to have another device in your house that has 5 or 10gbit ports so you could do some controlled testings. Things like speedtest are actually very complex with lots of software like web browsers involved.

 

[Victel]

All those commands mean is that the port electrically is running at 5gbit. Very technically it is always transferring at 5gbit. A ethernet port either transmits at full speed or zero. Those numbers are not real useful so most software measure data transfer over a period of time and take a average. The problem is there is something that is capping the throughput at a lower rate than the physical cable can do. A example would be say you run a vpn on your router. It might have gigabit wan ports and the ISP plan might be gigabit but the overhead of the encryption will put so much load on the cpu that you are lucky to get 30-50mbps on most router. Hard to say in the case of a pc since it is unlikely a cpu bottleneck. It almost has to be some setting or driver issue that is causing it to limit less than what the physical media can do.

Again this is why it would be nice to have another device in your house that has 5 or 10gbit ports so you could do some controlled testings. Things like speedtest are actually very complex with lots of software like web browsers involved.

I figured out what the problem was adguard's WFP Network setting. I turned it off and I'm now getting closing to 5g now. Thanks for the help