I have had Ultrafast 2 G.fast installed in my property for 2 weeks now.
Although the sync speed is consistent at 330-340 MBps download, I have osbserved that throughput variation is between 100-280 MBps.
I do the measurements whilst WiFi is turned off and all other wired devices are disconnected, with only the main PC connected to the Ultrafast Smarthub via Ethernet. All background tasks stopped when measuring, including anti-virus, dropbox, etc. I use BT Wholesale and Ookla tests to measure the speed. I have done a few measurements and charted them in spreadsheets, at different times of the day.
Why is this? Could it be an issue with the router?
There are no issues with upload and ping latency. Upload is pretty much consistent at between 30-40 MBps both Sync and Throughput.
Thanks @Mixer. But i think that Youtube video doesn't prove anything.
Why do you say the BT Wholesale Speedtest is not good?
Can anyone explain please. I am curious...
The BT Speedtest has been flaky for some considerable time. You can get different results depending on what browser you use , in my case using Firefox gives a speed 40% lower than using Chrome. The problem is that the test uses Flash which is no longer supported and all other test sites now run without flash.
Other problems involve running the further diagnostics test which doesn't always work or the test just doesn't run.
My main concerns are: (1) 'inconsistency' of throughput, and (2)'drop' of troughput from sync speed.
If it was only (in)consistency that i'm worried about, would the BT Wholesale Speedtest be at provide a good measure, consisdering all tasks stopped and only a single wired device connected to the router?
Is the Ookla test flaky too?
SpeedOf.Me bas been suggested which i will use when i get the chance... Are there any other speed tests worth trying, bearing in mind i'm looking for consistency and fair troughput results? which will help me decide to keep or not to keep Ultrafast G.fast.
I think I would trust Ookla more than speedofme.
Speedofme is very optimistic when running a test on Infinity 2, generally showing high 70s and bursts of 82.
Whereas on a Virgin 300 service that tests 386 down everywhere that supports 100+ Mbps speedofme consistently gives the results in the first image, a fraction of the speed attainable. Ookla consistently gives results like the second image. (+ or - 1Mbps on download)
The BT wholesale test can sometimes be quite accurate, as in agreeing closely with third party test sites, think it just depends on your luck/browser or what have you.
The Ookla test was one minute after the speedofme me test.
i would recomend https://www.thinkbroadband.com/speedtest
i know little of G.Fast but is it possible its just contention at the exchange or since its still in test stage maby they are experimenting with settings before it all go's live?
I would imagine there's perhaps a slight misconception regarding the published info. It is highly unlikely that G.fast would be released in the 'wild' as it were, if it had not proven itself during initial testing.