dslreports logo
    All Forums Hot Topics Gallery


how-to block ads

Search Topic:
share rss forum feed


Seattle, WA

help solve speed test mystery

I'm getting inconsistent results from the dslreports Flash speed test, in a very puzzling way...

I have two desktop computers in my office, both running Windows 7 64-bit, both connected to a NETGEAR WNR3500L router and a Motorola SB5100 cable modem, running Comcast HSI internet. There is a Vonage box (VDV22) attached downstream of the router, but no devices are connected to its ethernet out port. My primary machine (A) has a lot more software installed, while the other (B) is a pretty fresh Windows install - other than that they are similar in terms of CPU, memory etc. When I run dslreports Flash speed test, I get consistently different numbers between (A) and (B) ... (A) routinely gets download speeds between 6500-7500 Kb/s, while (B) reports much better scores: around 11000Kb/s. OK so that's puzzling (and I've tried switching ethernet cables, shutting off the firewall, etc) but wait ... it gets weirder...

If I go to speedtest.net and run the test there, I get very consistent speed reports between (A) and (B) - both in the 21-22 Mbps range. I'm not that bothered by seeing different numbers between dslreports and speedtest, but the fact that I'm seeing consistent and substantial differences between the numbers for PC (A) and PC (B) on dslreports but no such differences on speedtest ... makes me scratch my head.

I've only written download numbers, but upload follows the pattern - on dslreports (A) is much slower than (B), but on speedtest they're in the same range.

OK now I've tried one more test - at integra.net/speedtest, and it gives the opposite disparity: (A) returns download speeds around 22000kbps and (B) returns scores around 8500kbps (and in this case upload speeds are reported as very close between (A) and (B).

This makes me confused and frustrated, and more-so every step of the way. I can't figure out why (A) would be slower than (B) on dslreports, but then that result is thrown into question by (A) and (B) getting such similar numbers on speedtest.net, and then everything seems like a big can of worms when I look at the reverse numbers at integra.net.

What could be driving such massive and consistent disparities between the way the different sites report speeds on two different machines? And how can I come up with data that I can actually use to analyze/optimize my system?

(I've left the unit designations exactly as the sites report them: "Kb/s" from dslreports/Flash, "Mbps" at speedtest.net, and "kbps" at integra.net)