No buffer bloat improvement on XR500

[Nuschler22]

Thanks for any help provided in advance.

 

I have the XR500 with Gigablast from Cox (don't judge me, it's the only provider available.  When I do a normal test at speedtest.net, I average low latency to the testing sites (from 11 to 14 or so ms within my area, and excellent long range latency.  For example, I'm near 33ms typically to San Francisco which is about 750 miles away.   My speeds average about 500mb-750mb DL and 35mb UL fairly steadily from speedtest.net, which is fairly low compared to other speed test sites but seems more reliable and accurate.

 

When I do the network connection test in the benchmark, I always get readings that spike quite a bit for ping test under load.  I'm the only one using the internet when I game, or running these tests, so I'm at a loss as to why no settings seem to work to reduce it.  Speed is always A+, even when I'm at 80 percent restricted on DL/UL congestion control, but I always an A in both latency and ping under load with some strange spikes.  I've set congestion control to 20 percent, and it doesn't seem to matter.  

 

I only have congestion control enabled playing CoD on the PS5, and I use the PC for the tests.  I've added and enable the PC under the QoS settings under WorkAtHome, which I assume is meant for PC's, but I never get the reduced speed or testing.  I've assured that it's the correct IP address for the device and it's listed as "Desktop," but I'm not sure why it doesn't work for testing.

 

Just curious what I'm doing wrong, and how I can decrease buffer bloat.  Thanks.

[Netduma Fraser]

We're still improving the results for Benchmark so that will likely explain what you're seeing. Benchmark also uses the Google servers for Ping and Ping Under Load so your connection to these servers may not be as good as other servers you're testing but it wouldn't impact your game necessarily. 

Ping Under Load is only an issue when your connection is saturated (i.e. all your speed is being used), with just the one device playing you would never experience this. WorkAtHome is for prioritizing Zoom, Skype etc so not necessary for gaming. When you prioritize this supersedes the Congestion Control percentages so it's likely those programs would use port 443 which would then be prioritized and this is also the main point for web browsing so when you do the test it would get prioritized as well in that scenario.

[Bert]

If it's cable internet, there is often an increase in latency when your connection is loaded. It's due to the modem and you can't fix this with a router, only limit your speeds so it occurs less. You often have to get very low to eliminate this, below 50%.

[Nuschler22]

On 4/8/2022 at 4:15 PM, Netduma Fraser said:

We're still improving the results for Benchmark so that will likely explain what you're seeing. Benchmark also uses the Google servers for Ping and Ping Under Load so your connection to these servers may not be as good as other servers you're testing but it wouldn't impact your game necessarily. 

Ping Under Load is only an issue when your connection is saturated (i.e. all your speed is being used), with just the one device playing you would never experience this. WorkAtHome is for prioritizing Zoom, Skype etc so not necessary for gaming. When you prioritize this supersedes the Congestion Control percentages so it's likely those programs would use port 443 which would then be prioritized and this is also the main point for web browsing so when you do the test it would get prioritized as well in that scenario.

I only have the one device (PS5) on, but when I do the test I obviously just have the PC running for it with nothing else connecting.  So, not sure why I'm getting a higher ping under load.  I don't use Zoom, Skype, etc.  

[Nuschler22]

On 4/16/2022 at 6:55 AM, Bert said:

If it's cable internet, there is often an increase in latency when your connection is loaded. It's due to the modem and you can't fix this with a router, only limit your speeds so it occurs less. You often have to get very low to eliminate this, below 50%.

Thanks, appreciated.

 

My speed is set to 500mb for DL which is about 2/3's of what I normally get, and I have it at 50 percent on both.  I'll give 25 percent for both a try and see how it works. It's still something like 125mb DL and 8mb UL which should be plenty regardless.

[Netduma Fraser]

28 minutes ago, Nuschler22 said:

I only have the one device (PS5) on, but when I do the test I obviously just have the PC running for it with nothing else connecting.  So, not sure why I'm getting a higher ping under load.  I don't use Zoom, Skype, etc.  

We need to improve the results on Benchmark so you can pretty much disregard that as it would unlikely to be that bad and at your speeds it is unlikely to become congested to where you see bufferbloat, especially with QoS active.