D1NZ:Otagolad:D1NZ:insane:Otagolad:SneakerPimps: I still haven't contacted Snap, but it seems as if it's improved. At least I can now max out my line with multi-thread.
So Snap must have found something with Dummyemail7 and applied it to the rest.
I'm not sure there has been any improvement for me - on 200/200 UFB the best speed I can get to the Dreamhost Los Angeles server is 6.93 Mb/s down and 3.74 Mb/s up - this is quite a difference from D1NZ's test of 81.65Mb/s down and 53.89Mb/s up and we are both in Christchurch on the Enable network and with Snap.
Even just the slightest amount of packet loss at that distance would make a massive difference. Would be good to see a bandwidth graph while your test is underway to see how smooth the tcp ramp up is.
Could it be hardware related? say like router or the fibre backhaul?
From my end it is unlikely to be hardware related as my setup is a home build computer with Intel Core i7-5960X Haswell-E 8-Core 3.0GHz processor, ASUS ATX DDR4 3000 (O.C.) Motherboard X99-DELUXE, 32GB Ram connected to an Asus RT-AC68 router by Cat 6 ethernet cable which is connected by Cat 6 to my Fritzbox (wifi disabled).
D1NZ - are you able to run another speedtest today to the LA Dreamhost server and post again to show it wasn't an anomaly?
Also I don't run any of those All-In-One router, I uses a self build PC/server and running pfSense, my build as following:-
Intel Xeon E3-1225 V2 @ 3.20Ghz
Intel S1200KPR server board mini-ITX
A-data ECC 4GB x 2 1600MHz
Intel Ethernet Server Adapter I350-T2
Samsung 840 Pro 120GB
Thanks for doing that so quickly - also, do you think that using your own equipment would account for such a massive difference in speed tests?