100M New BT speed broadband speed test - Page 2

Chrysalis · 02-04-2012, 20:49

23.17

26.24 on speedtest.net

yes my downstream performance has improved latetly mostly manages to get over 20mbit now (not same of upstream tho that getting worse).

Blairhoyle · 03-04-2012, 07:45

I don't trust these sites at all.

JadeFalcon · 03-04-2012, 09:30

as a true test i always use something like Utorrent and download a linux iso such as ubuntu there's always a ton of seeds on it... on speedtest.net and others i dont seem to get over 30 on my 60meg connection but i can get the ubuntu iso at 7.4MB/s

i think this is why virgin get complaints about not reaching advertised speeds is because they think just because they have 60/100 meg, that every site on the internet should give them full speed, when most wont give that to one connection. dont work like that, thats why i always use a torrent to test available speeds

Chrysalis · 03-04-2012, 12:48

speedtest.net is multi threaded tho so you have a bad configuration somewhere if you need dozens of threads to max out or it could still be down to congestion as higher threads can hide congestion.

Nidge41 · 03-04-2012, 15:59

Gave me a nice 103MBPS for some strange reason

craigj2k12 · 03-04-2012, 16:17

should give about 105 providing theres no congestion, a bit more on the xxl package

roughbeast · 03-04-2012, 16:35

Quote:

Originally Posted by JadeFalcon

as a true test i always use something like Utorrent and download a linux iso such as ubuntu there's always a ton of seeds on it... on speedtest.net and others i dont seem to get over 30 on my 60meg connection but i can get the ubuntu iso at 7.4MB/s

i think this is why virgin get complaints about not reaching advertised speeds is because they think just because they have 60/100 meg, that every site on the internet should give them full speed, when most wont give that to one connection. dont work like that, thats why i always use a torrent to test available speeds

Torrents won't give you a good estimate of speed even when they are up to their max. Your download and up load rate are too dependent on file availability and other people's connections. You would be better off downloading multiple files simultaneously from a know-to-be-good server, then combining their download speeds.

http://gamefiles.virginmedia.com/blueyondergames/

You will need to download up to 8 large files to max out a 100Mb connection. You will get about 2MBPS from each file.

craigj2k12 · 03-04-2012, 17:21

Quote:

Originally Posted by roughbeast

Torrents won't give you a good estimate of speed even when they are up to their max. Your download and up load rate are too dependent on file availability and other people's connections. You would be better off downloading multiple files simultaneously from a know-to-be-good server, then combining their download speeds.

Why not? as long as your outside protocol shaping hours it should give a VERY good picture

buckleb · 03-04-2012, 17:25

My connection is shown as a pretty consistent 86Mb/s (110Mb product) when tested on that BT site, though it's better than the usual Speedtest result of 30Mb/s.

However, downloading from a fast server (using a download manager) gives me the full-on 13.69MB/s.

RealDiamond · 03-04-2012, 17:31

Linux distros have High speed servers behind their torrents not just domestic users. (Linux may be open source but has Millionaires funding its PR and websites)
Ive had all 10Mb/s (1.2MB) filled with just 1 peer/connection.
Compaired to other torrents which have needed over 20-30 connections to max out my download.

new site test failed.
speed 2.5Mb/s test locks up.
Speedtest.net
9.987Mb/s

roughbeast · 03-04-2012, 19:58

Quote:

Originally Posted by craigj2k12

Why not? as long as your outside protocol shaping hours it should give a VERY good picture

With torrents, if you are trying to measure your download capacity, it's nothing to do with your traffic shaping or your connection quality. It is to do with the availability of the files your downloading across a number of other people's upload capacities. Granted, that if you get a really good torrent with hundreds or thousands of seeds and few leeches, like you get with JayBob, you will get really quick downloads. However, even before I started to use BTGuard, (proxy via Canada) I never maxed my connection.

craigj2k12 · 03-04-2012, 21:22

the only reason torrents were 'invented' was due to the immense speed and zero redundancy, your reasoning is rubbish

roughbeast · 03-04-2012, 22:10

Quote:

Originally Posted by craigj2k12

the only reason torrents were 'invented' was due to the immense speed and zero redundancy, your reasoning is rubbish

Immense potential speed, rarely achieved, over typical torrent swarms, sufficiently to max out a 100Mb connection.

This is where we came in. I am merely contending that multiple file downloads from a reliable fast server is a better way of testing your full download potential than almost any other means.

We can argue forever and a day about torrents. I swear by them for downloading content I want quickly, but not for speed testing purposes.

Skie · 03-04-2012, 22:34

Torrents main benefit is that they quite effortlessly distribute the bandwidth across the users rather than have it all come from one bottleneck. Each user does their own bit. It was never about speed, just removing that single point of failure and allowing anyone to distribute large files. Heck, Cohen built it because he could.

And I'd actually expect the superhub to have problems maxing out a torrent that wasnt being seeded by someone with a large pipe. If the average seed speed is low it probably can't handle connecting to over a thousand seeds to achieve the line maximum.

craigj2k12 · 04-04-2012, 00:18

it was never about speed, but your able to download a several gigabits providing you have the capacity and computer up to the job, discounting the fact that the superhub probably cant cope with more then 10 simultaneous connections yes, I see your point, but in modem mode, using a proper router, I would expect torrent downloading will always give far more accurate results than a speed test

Youve also got to remember, which has been mentioned before, VM dont provide the capacity to download at high speeds from everywhere, we all know London speedtest.net server gives better results from VM connections than say from maidenhead, however maidenhead can give results over 600mb from other ISPs, so its a VM problem not the servers fault. I expect every server in the UK on speedtest.net is way more than capable of measuring at 100mbit, the bottleneck lies somewhere with VM