dslreports logo
 
    All Forums Hot Topics Gallery
spc
Search similar:


uniqs
410
oooeemrlee1
Premium Member
join:2003-12-10
La Quinta, CA

oooeemrlee1

Premium Member

New Speed Test - Location error

Per your request, the new speed test estimated my closest server to be on the East Coast. I'm located in So. California so the speed test is not valid. FYI.
»[Cable Speed test: 7.29/5.05 95 ms]

justin
..needs sleep
Mod
join:1999-05-28
2031

justin

Mod

Thanks for noticing that.

Try it again, there was a temporary glitch with the selection of the fastest server.
tispectrum
join:2001-03-17
New York, NY

tispectrum to oooeemrlee1

Member

to oooeemrlee1
I'm on the east coast, and the speed test is from San Francisco. On your Java test my download was about 75mps. On 3 other test sites it varies between 65 and 108 mps. On your new speed test, albeit from SF, it was about 25.

justin
..needs sleep
Mod
join:1999-05-28
2031
Billion BiPAC 7800N
Apple AirPort Extreme (2011)

justin

Mod

thanks for the feedback.
For my reference the link to the results is: »[Cable Speed test: 21.31/4.42 95 ms]

First, you are on the east coast yet your ping time to the east coast server (this site) is 68ms, which is frankly pretty high. It appears to go all over the country.

But the SF distance is even further, 97ms.

The reason it picked SF is because:

0 latency 94.5ms from San Francisco, CA, USA
1 latency 269ms from Sydney, NSW, Australia
2 latency 220.5ms from Dublin, Ireland, EU
3 latency 117.5ms from Michigan, USA
4 latency 133.5ms from Charlottesville, Virginia, USA

the latency at that one point was lowest to SF. Although this isn't a very good measure as it only does one quick check.

Now, downloading from SF, the best it could manage was
21 download completed in 6879 ms latency 555 length=18320000 mbit=21.31

so the speed is what it is, to Amazon in SF.

If you run it again, probably it will pick Michigan, and then it should be the same as the java test, for you?
Mele20
Premium Member
join:2001-06-05
Hilo, HI

1 edit

Mele20 to oooeemrlee1

Premium Member

to oooeemrlee1
I ran this test a few minutes ago and it chose the "best" server (San Francisco) and the results were accurate. I just did it again now and it chose the East Coast server and did not show ping time for San Francisco although it said it checked three servers. Test results were not terrible for using a server so far away with horrible ping times but still...we should be able to choose the server. What if I want to test to a server it did not choose? This test should not limit user choice.

Almost 300ms ping (chosen as the lowest ping of the three) is ridiculous. Yeah, this second test was done on a Hyper-V virtual machine running Win 10Preview and the first test on was on the host machine Windows 8 Pro but that shouldn't make a difference. Maybe there was a momentary glitz causing the high pings....I'll have to do more tests to get a better idea of accuracy.

It only likes IE 11 on Window 10 Preview build 9926 with low pings and accurate results. It apparently doesn't like Fx 31.4 ESR which is the LATEST ENTERPRISE version fully supported by Mozilla. Very high pings for all test sites and it chose Michigan. I know the test says current browsers but this IS the current Fx Enterprise browser and it should work better with this test.

It also doesn't appear to like Pale Moon 25.2.0 9x64) where the test fails repeatedly. It tests fine with the same Pale Moon on Windows 8.

justin
..needs sleep
Mod
join:1999-05-28
2031
Billion BiPAC 7800N
Apple AirPort Extreme (2011)

justin

Mod

Can you copy a link to the results of your test.
Just click on the blue date, and paste that URL

I'm going to make the latency a bit more of an average of 2 or 3 goes, not just one attempt.

Yes there will be a way to select servers etc however first I'm concentrating on making it trivial to use but accurate.
Mele20
Premium Member
join:2001-06-05
Hilo, HI

Mele20

Premium Member

Which test?

In the order I did them:

»[Cable Speed test: 32.90/5.29 99 ms]

»[Cable Speed test: 25.60/4.53 286 ms]

»[Cable Speed test: 29.40/5.30 78 ms]

»[Cable Speed test: 32.70/5.23 120 ms]

»[Cable Speed test: 32.80/5.33 92 ms]

»[Cable Speed test: 28.80/4.57 199 ms]

The second one and the last one are the two that did not choose San Francisco.

Good if it has several attempts for latency and chooses the best.

Glad to hear eventually we can choose the server.

Will there be a way to post the results to our ISP speed table?

justin
..needs sleep
Mod
join:1999-05-28
2031

justin

Mod

what is our ISP table?

yes all results are exhaustively kept and there will be ways to see them when I have enough evidence that they are reliable.
Mele20
Premium Member
join:2001-06-05
Hilo, HI

Mele20

Premium Member

This:
»/archi ··· 02&r=827

The java test here to Amazon server is terrible for speed (at least this one time I tested just now).
Mele20

Mele20

Premium Member

Click for full size
The Flash test is accurate for download and a bit high for upload ( I have 30/5 speed tier) on Comcast in Denver. It doesn't offer to let me post the results to the RR speed results table. Java test lets me post that very poor speed in the table.

Flash test using Amazon cloud is WILDLY OFF on download speed (claims 52 mbps where I have 30). Upload is accurate.

I wish we still had nac.net Java test. It was the best over all these years.
oooeemrlee1
Premium Member
join:2003-12-10
La Quinta, CA

oooeemrlee1 to justin

Premium Member

to justin
Justin, thanks. We have network problems out here. Tech checking nodes between me and the Fiber plant tomorrow.

justin
..needs sleep
Mod
join:1999-05-28
2031
Billion BiPAC 7800N
Apple AirPort Extreme (2011)

1 recommendation

justin

Mod

said by oooeemrlee1:

Justin, thanks. We have network problems out here. Tech checking nodes between me and the Fiber plant tomorrow.

Are you you supposed to be getting better than those 30/3 to 30/5 results?