CA3LE

I just ran the test in Windows on Firefox and task manager is showing the same thing as TMN.

 

You're getting up to nearly 450 Mbps results in the upload test, task manager should reflect this.

 

Could it possibly be going through a different interface connected on your computer?  Maybe another ethernet connection or wifi?

 

... with a connection like that, you'll have to give my new beta a spin.  The current upload test is limited to 100 MB size, and single thread.  The multithread upload test is only in the beta and is the core of why I built an entirely new version.  Invites for the beta a now closed but you're going to get an invite automatically because of your rank here.  Looking to send that out soon.

 

Here's xs1's final upload result on Frontier

TestMy.net Test ID : qX0Ybzziu

 

And here's what he was able to pull using the current version. 

 

On 10/13/2022 at 7:18 PM, Pheelix said:

I think testmy.net needs to update there testing process.

 

Hey Pheelix, cool name btw.

 

I notice you're not signed up for the beta.  If you want to see the updated process before the general public vote on this topic >> https://testmy.net/ipb/topic/34612-beta-testers-welcome/

 

 

Voting will close soon and the team will be built.

I have Xfinity and my modem rarely resets like that.  When it does it's usually around midnight or 2 am.

 

If it's doing that frequently it's a sign of some kind of issue.

 

couple questions...

• Where is this modem/gateway?  Does it happen to be in a confined space, possibly running hot?
• I know the gateway has restarted recently but did you happen to do a full power cycle?  (unplug for 10 seconds)

 

7 hours ago, spudler_t said:

If I owned my own Gateway would they still lie and claim they need to do these updates (just restarts a majority of the time) and @CA3LE do you feel it is worth it to buy your own gateway as they are not even remotely cheap.

 

... again, it shouldn't be restarting frequently.  Once a month at most.

 

Every so often you will require updates from Comcast and sometimes they may require the modem to restart.  No matter which hardware you use and regardless of the ISP.

 

 

I feel it's always smart to own your equipment.  A nice quality Netgear gaming router and Arris Surfboard modem compatible with your package is what I recommend.  Long term it saves money.  When you want to upgrade having a separate modem and router makes it less expensive... and gives you the most choices and options.

 

Right now I'm currently renting a gateway.  I use it only as a modem then it goes into my own router.  Personally I'd never trust them (or anyone else) with my networking.  I had to rent their modem to get unlimited data.  Now they allow unlimited with your own modem but for +$30 /month, vs +$25/month (inc gateway rental).

 

But if you don't have or need the unlimited you'll save $15/month.  I see SURFboard S33 (check with Comcast and make sure this works with your package first, it should) for $100-$120 refurbished on amazon.  So 7-8 months break even.  Then for a router I might get a Netgear Nighthawk Pro Gaming XR500 refurbished for $105-$150. Let's say 8 more months.  So 16 months and you'll be ahead.

 

That router is limited to 1 Gbps ethernet but I think the over all quality of the connections will be amazing.  There are 2.5 Gbps ethernet options but at a premium right now.  Netgear RAX120 has 1x 1/2.5/5Gbps and I've seen those as cheap at $149.

 

It may be tempting to get a modem/router combo, like the Netgear CM2000 for $189 refurbished.  But if anything goes wrong, it usually costs more to replace.  And then you lose that upgrade flexibility I mention above.

 

For this type of gear I would definitely save the money and get refurbished or used.  If it works for a week or two, chances are it will chug away for years.

The confusion is because of the difference between Mbps and MB/s.

 

Mbps or Megabits per second is measured in bits. MB/s or Megabytes per second is bytes.  There are 8 bits per byte.

 

So to convert MB/s to Mbps you times by 8.

 

So your seeing about 2 MB/s .. which is the same as 16 Mbps. 

 

Bytes are used when you're talking about binary file size or file transfer speed.  Bits are used to express connection speed or thruput.  .... e.g. I downloaded a 14 MB file at 2 MB/s, so my connection speed is 16 Mbps.  I can stream at 16 Mbps.

 

Hope this helps.  Happy Testing!

Give us some more details on your network.

 

What router and modem do you have, how many computers and are they all wifi?

 

Have you tried testing directly connected, from the modem to one of your computers?  That will help us understand if it's something within your network or if it's coming from Vodafone.

Hi Steve,

 

Can't really make them.

 

You definitely have something happening.  Appears like sometimes you're fine, then other times you get that seesaw effect.  The dreaded intermittent issue, worst kind to nail down!  How often does this happen?  

 

Have you tried any other browsers or computers?  Are you fairly certain that it happens in all browsers and on any device connected to your network?  Or have you only testing in one browser on one computer?

 

7 hours ago, Catbear said:

Too bad I don't have my own server, but I'll give it my all on my local network! 💪

 

You won't need a server.  You can install Docker on Mac, Windows or Linux.  Then it's just a matter of a couple easy commands and it's running.

 

7 hours ago, Catbear said:

Together, we'll conquer the internet!

Let's rock this Beta! 🚀

 

You're getting me all psyched up!

Catbear 🐱🐻!  You're awesome!

 

 

Please join the beta to get early access >> https://testmy.net/ipb/topic/34612-beta-testers-welcome/ 

 

I think you're going to love playing with the new tests on that connection.  I'd like to see how it runs on your local network too.

 

 

 

9 hours ago, Catbear said:

I'm stoked to share my experiences

 

I'm stoked to read them.

 

Welcome to TestMy.net!

 

By the way, sweet connection!

Welcome @Little Banshee!

 

TestMy.net is currently only testing your single file transfer performance (on upload).  This is close to what you'll see if you try and upload a single file somewhere.  This also shows you what kind of streaming quality your connection is capable of.  If you tried to stream over 70 Mbps, you'd probably have issues.

 

My new beta is built around a multithread upload test.  Join us in testing >> https://testmy.net/ipb/topic/34612-beta-testers-welcome/

 

The multithread upload test will help you see how fast you are when you combine multiple transfers together.  Comparing the two tests helps you understand your connection better.

You can use any OS, any browser, A/V, firewall etc.

 

Usually any computer that has a fresh install will perform well.  It's only after introducing variables that limit or get in the way of the connection that there are ever issues.  I have dozens of programs installed on my computers and none of them get in the way of my normal transfer, ever.  Nor do I ever encounter programs or settings that affect my TMN results, ever.

 

There are no special requirements.  If you run the test on a memory constrained machine, a machine with something misconfigured or with software parsing your browser data... then yes, you're results will suffer.  That's why we test!!  But that doesn't mean you have to make stars align.  Just if you start with crap you'll usually end up with crap results.  If the conditions improve TMN results always improve.

 

Now, we know that tablet isn't crap so there must be some kind of crap causing it to look like crap.

 

2 hours ago, MFP said:

The Mod, who I was not talking to said, "Seems like you're always ready to attack me. "

 

... what?  Dude, I'm the Webmaster and clearly you were writing to me.  If you're writing TestMy.net, you're writing me.  I am TestMy.net.

 

Back to my post from Friday, have you tried resetting Edge to default?   ··· >> Settings >> Reset Settings >> Restore settings to their default values

 

TestMy.net is only trying to help you by showing you what it sees.  If two browsers on the same machine are performing 10X different, there's good reason you're getting those results.  It doesn't just happen.

Both engines are proven to perform well on TestMy.net.   As long as they're running correctly.  I run them both regularly, every day.  On different operating systems, different machines all over the place.  I never see what you're seeing unless there's good reason.

 

Seems like you're always ready to attack me.  If you don't like my test, my results or the advice I'm giving then why are you here?  

 

11 hours ago, MFP said:

Sorry to burst your bubble

 

Don't worry, you didn't burst anything.  I realized before this even became TestMy.net over 20 years ago... I'm never going to please everyone. 

 

Obviously you can't be helped, you may not even know how to accept help.  When people take a lot of time to help you, you spit on their advice.

 

If I wanted spit in my life I'd work in a prison, zoo or insane asylum.

 

Some people act like I owe them something.  This site is given to you AS-IS, take it or leave it.  If it works for millions of other people for over 20 years... but doesn't work for you then you should probably wonder what's different about your setup than everyone else.  Instead you join the camp of those who shoot the messenger.

 

The other people here testing with similar Chrome setups but getting much better results prove that the issues are unique to you.  But you want the problem to be someone else.

 

Your wife probably clicked something and got the device infected, being an analog girl and all.  It could infect the base OS browser but not other installed browsers.  I would't ignore those results, they only happen for good reason.  When I've played with surface tablets in the past, I've been blown away how well they perform on TMN.  Using Edge. 

 

Edge ≠ Internet Explorer.  For the most part it performs as well or even better than Chrome in my testing.  I no longer have to program specifically for Microsoft's browser, as it now respects the standards.  Any browser that accepts the standard will perform well on TMN, which (after Edge switching to Chromium) is every-single major browser.

On 7/7/2023 at 3:24 PM, MFP said:

So I fired up Firebox and my linear tests to Dallas went from 30 to 230. (ident 04) No reboot, no changes, no nothing. Just a different browser.

How do you explain this?

There are still way too many variables here!

...

Meanwhile if you could answer my burning question.

 

 

Hi MFP, after my last response I got back into development again.  I hope it all makes sense when you see the beta.  Very proud of how it's turning out.

 

I didn't notice this post until just now... when catching up on the topic I was focused on the post after the one I quote here.

 

When that happens, where one browser is much faster than another on the same machine, it always points to software as the culprit.  There's no other reason why that would happen.

 

I would reset that browser to factory defaults to see if it improves.  But first...

 

Often it can be some plugin you've installed.  You can try systematically disabling them to find the culprit.  I would disable all of them first, if speed improves THEN disable them one at a time until you find it.

 

Routinely installed software could have added the plugin/extension.  Or maybe a plugin you've personally installed, may have even ran fine until an update. 

 

A plugin that has the ability to parse the HTML content of the pages you're using, like a deal finder or ad blocker can severely cripple your browser's performance.  Especially if the author of the software is doing something nefarious.  Which at least to some degree pretty much all of them are.  (At least anonymizing your data and sharing it with third parties.)

 

TestMy.net's Linear test is able to detect this when other's aren't because the data is streaming.  It's like you're reading a document from top to bottom.  You can't split the data as it's flowing to hide what's really happening.  If something is parsing the data as I'm sending it to you it's possibly going to slow down my transmission to you.

 

The plugin may be harmless and just not a very optimal program.  Or it may be parsing your data for evil.

 

If the plugin is parsing for a few keywords it might not be noticeable but depending on what it's trying to accomplish and how in depth it's parsing it might really slow things down.  You may not even notice in normal browsing with fast enough internet... but TMN will notice.

 

Only way I'd ever ad block is with DNS, only way I'll look for deals is with good old research.  NEVER EVER EVER EVER install ANY browser plugins, EVER.  There is no exception.  If you absolutely need a plugin to do something legit (like metamask) then virtualize a separate desktop or have a specific browser just for that task.  If something needs a plugin, you're most certainly giving it too many permissions.  I trust no plugin.  I could audit their code today and they update it all tomorrow.

 

Hey, maybe it's something else.  I'm literally 1000 miles away.

 

Back to reset that browser to factory defaults; It could be a simple setting in the browser.  Might be buried deep in advanced settings where it could be nearly impossible to find.  If all else fails, just reset the browser.  All browsers these days seem to have some option for that.

 

If it is a specific plugin or piece of software and you narrow it down, please let us know what you find.  The examples I gave above are real, from many TMN members in the past who've had similar issues and reported back their findings.

 

Just remember, it's not happening for no reason at all.  TMN doesn't lie.

 

I started writing more here and then started another topic.

 

 

My beta gives anyone the ability to run TMN on any computer they'd like.  Run it locally on a Docker images I've prepared for you or install it on any web sever, with or without SSL.  It only takes 2 files, a total of 1800 bytes of code on your end.  What speed will you see between two wired 1 GbE?   Exactly what you'd expect to see, watch for yourself.

 

The Mac and Windows machines are fairly distant, non-direct routes.  The connection traverses 2 switches, then the router, then another switch also adding in about 100 ft of cable before it arrives.  Not lab conditions, I wanted all of those real world variables. 

 

If you inspect I think you'll agree, TMN's results are pretty exact. 

 

tl&dr

 

So average 916 Mbps | 904 Mbps.  Adding 6% network overhead you get to 971 Mbps | 958 Mbps.  Which is right in line with what we see coming across the interface in reality.  

 

I did the same LAN testing when developing the current version (v18) you're using now, always do.  But now I'm giving the ability to do it yourself, super quick.  ... I've also done that before but not like this.  This is very different.  Cut and paste a few commands into Terminal or PowerShell and you're running local tests on all your devices in a few seconds.  You can also audit all of the code and understand the entire container in just a few seconds.

 

Pretty excited to share that with you all.  Curious how people will use it.

Sorry I dropped off, like I said, was getting back to writing the beta.  It's taken on a life of its own, controlling me into building more.  In a good way.  Never enough hours though.

 

If I think of anything that might help more I'll update this thread.  Please let us know if you figure out anything solid on your end.

Reboot the computer, router and modem and give it another try.

 

Testing with the same options as you and the test runs as expected so I don't think there are issues on my end.

Currently the multithread option only works on download tests.

 

The new beta multithreads upload, https://testmy.net/ipb/topic/34612-beta-testers-welcome/.  Hasn't been released yet but I'm getting it ready.

 

I suspect you may get faster speed in the multithread upload test.  Other speed tests multithread by default.

 

So you may be able to get > 300 Mbps upload but that's when combining multiple uploads together.  You'll probably only be able to upload a single file at the speed TMN is showing you. 

 

It's all in how you look at it, I'd personally call that 52 Mbps not 300 Mbps.  My new test will help you see the difference.  Most people just want to see the big number but depending on the circumstance the same device can have a faster result either way.  I see the linear test blow away multithread on my 4G connections sometimes, other times it's the other way around or it's even.  Sometimes the extra connections help and sometimes they can hurt.  So I think it's always best to test both ways.

 

52 or 300 Mbps, actually both awesome results in that test in my opinion.  You'll definitely score much higher with the new test.  Vote on that topic and I'll send you an invite when it's ready.

 

I believe your ISP should deliver 300 Mbps in a single upload.  If you're showing 52 Mbps then it's safer to say that you have 52 Mbps.  You can stream up to that level, not 300 Mbps.   If you see this across all computers on your network it may be a limitation of your ISP.  Can also point to local network issues, we've even seen modem swaps fix issues like this.  People will get like 10 Mbps upload on TMN but normal speed on all other tests.... they swap out the cable modem and TMN reports normal speed again, back in line with the other tests.

 

Cases like those prove that other speed tests are a waste of your time.  They've proven over and over that you really can't trust them to help you.

 

My goal is to help you.

 

I agree, if it's across so many devices you should look to the common denominator.  

 

This is why I suggested that you test directly wired to the modem.  You said you already did this but when?  How were you testing?

 

If you can show the ~40 Mbps linear result on a computer with a fresh OS install, directly connected to the modem it will reasonably prove that something is either wrong with the modem, connection to the modem (cable or adapter) or Spectrum.

 

We have to eliminate all those variables, have you tried that combination yet?

 

50 minutes ago, MFP said:

I can get 200 from a computer that just got 40 as well. I can get 200 or 90 and then 40 an hour later. A one time snapshot doesn't prove anything.

 

You may be able to get 200 on one that drops down to 40... but did you ever get 40 on the fresh install (win11)?  From what I saw, it looked like that machine never showed a drop like the others.

3 minutes ago, MFP said:

To prove to you and others that it is not my computers or router.

 

It doesn't prove anything if the test is only capable of seeing things through a distorted lens.

 

It can be a bunch of different issues that we've proven over and over are not detectable with other speed tests.  

 

Something is preventing you from scoring well, it's not TMN and I don't really think it's Spectrum.  It appears to be a localized issue.  Sorry that's not what you want to hear.

 

Your speed improved when you booted the fresh win11 install, right?  I think something you have installed or some setting is messing things up.

other speed test ≠ testmy.net

 

They're testing different and don't look at the result the same way as I do.

 

You say you're having problems.  Why would you trust the other test's results ever after they've shown it working correctly when it obviously isn't?

1 hour ago, MFP said:

Problem with the CO test sever or Spectrum?

 

Not the server.

 

890.4 Mbps [111.3 MB/s] | Downloaded 200 MB in 1.8 seconds
Pre-Test: 735.2 Mbps [91.9 MB/s] | Downloaded 60 MB in 0.65 seconds
Download Graph: [504,656,702.4,845.6,735.2,763.2,497.6,610.4,692,890.4,936,931.2]
Test Latency 18 ms Avg - 10 ms Min
Latency Graph: [21,21,16,16,19,14,14,19,27,17,26,10]
Locale: co.testmy.net
Graph & Validate: https://testmy.net/x/tmn?resultID=SWZWDFC8O

 

481.6 Mbps [60.2 MB/s] | Downloaded 138.6 MB Linear in 2.3 seconds
Pre-Test: 375.2 Mbps [46.9 MB/s] | Downloaded 60 MB in 1.28 seconds
Download Graph: [151.2,194.4,236.8,375.2,450.4,487.2,478.4,475.2,482.4,488.8,481.6,478.4]
Test Latency 27 ms Avg - 16 ms Min
Latency Graph: [19,26,19,22,16,19,16,27,24,82,25,26]
Locale: co.testmy.net
Graph & Validate: https://testmy.net/x/tmn?resultID=Lh4JaCfuA

 

1 hour ago, MFP said:

It's not the computer (Samsung tablet) as it will do fine to the Spectrum speedtest beauty sites.

 

Comparing apples to oranges.

1 hour ago, CanadianHACKr_eh' said:

🤙🇨🇦

 

Looking forward to having you on the team.  I'll let you know when I'm ready.  

I thought that seemed a little high.  10 Mbps (10,000 Kbps) you can easily do.

 

On 6/28/2023 at 2:33 AM, spudler_t said:

I may be confused but isn't 1200/40 equivalent to 120000/40000 in kb/s??

 

So yes, 40 Mbps is 40,000 Kbps ... kb/s is also correct.  I just thought you meant kB/s because I abbreviate it as Kbps.  My bad.

 

If you wanted to, you could easily do 2 streams at 10 Mbps.

13 minutes ago, MFP said:

This includes me.

 

Includes most people.  Includes me as well.

 

Pulled up CenturyLink (really the only other ISP in the area) to my address and here's what I can get.

 

lol.  Wow!  You'll even give me unlimited data! 

 

You may want to say monopoly.  But it comes down to the infrastructure.  Cable laid the infrastructure with superior technology far earlier than telco companies.  Coaxial cable even in it's early stages has always been amazing.  It was immediately ready for high speed internet before it was even a thing.  The architecture of cable systems beat twisted pair, they laid more infrastructure easier.  They won for good reason.

 

It may be a monopoly that they don't have competition at the ground level.  They may just buy up anyone who tries to compete.  But really as a company today, are you going to come into a rural area and fiber up everything?  Probably not... but cable had laid the infrastructure early, in the analog days when hardwired made even more sense.  Today, they're going to want to beam you the connection. 

 

If I have the choice between wireless, old crappy telco tech and cable... I'll always choose cable.  That's why my nick is CA3LE.

 

Hell, I'd love me some fiber, just have never had those choice.  If I still lived in Phoenix I'd have a bunch of choices these days.  

 

So my options are 60 Mbps for $50 from CenturyLink or from Comcast...

 

Even if all of those packages perform at 50%... still better than CenturyLink.  At least in my area.  Now CenturyLink Fiber (FTTH) ... I imagine is a totally different story.

9 hours ago, MFP said:

You originally came off to me as a guy who knew a lot and has been doing this a long time.

There is a lot of big talk, hubris, and bragging by the author of this website.

All I have discovered with your site is what is leading me to question the bandwidth and rigor at your test servers.

 

I love when I spend hours upon hours helping someone, reading into all their tech support emails, taking hours to work someone else's problem in my head, explaining things in intricate detail, writing custom guides specifically for that person and taking time out of my family holiday weekend to quickly respond back... only to be attacked.

 

Talk about biting the hand that feeds you.

 

I understand your frustration and I've been trying to help.

 

It's pretty common that people attack and blame me and my site, I'm used to it by now.

 

Quote

(from MFP's email)

1.      How is it that you can have a 2000Gbps backbone, as you say on your website?

2.      How many IXP's are you connected to?

3.      Aren't most of the 43 or so US IXP's limited to a 100Gbps backbone?

4.      How can you afford all this bandwidth? I do see some ads when I turn off blockers.

5.      I now have test data from six of my computers, with up/down, multithread on/off, to three test servers, how do I summarize this to present to Spectrum? Spreadsheets are too complex.

6.      What are the reasons for such a wide variance between multithread on and multithread off?

7.      Why do I have to turn off Privacy Badger in Firefox and Brave Shield in Brave to get your tests to work?

 

Here's what I actually say on the website right now,

• Total network bandwidth capacity in excess of 2000 Gbps
• TestMy.net's servers are all gigabit or multi-gigabit uplinked
• Hosted network features multi-homed connectivity with bandwidth from independent Tier 1 carriers, combining multiple 10 Gbps connections to create one of the industry’s fastest networks.
• Global network reach utilizing over 25 Tier 1 network providers including Comcast, Cox, Time Warner, Charter, Qwest, Google, Level 3, Internap, NTT America, Equinix & Telefónica to name a few
• Extensive peering relationships across North America, Europe and Asia

 

That was actually written so long ago, needs to be updated and includes much more currently.  After I get the beta released I'll probably start updating text across the site.

 

Total network capacity.  This is the network capacity of the hosts I choose.  I choose hosts who peer well with other hosts and have in many cases direct connections to the larger ISPs.  Those are their peering relationships.

 

Quote

How can you afford all this bandwidth? I do see some ads when I turn off blockers.

 

I've definitely had to be creative in the past to afford to run this site.  From 1999 to 2001 I had to shut down because I exhausted all of my options.  And then lost all the code.  Long story about how it was able to come back.

 

TestMy.net today I have a cabinet of servers at Colohouse (formerly Data102) about 15 minutes from my house in Downtown Colorado Springs CO.  This is where the (currently) centralized databases run.  When you visit "testmy.net" it's served from Cloudflare CDN and goes to my rack here in Colorado.  When you use the Colorado test server, you're hitting a VPS running in that rack.  My cabinet is located only about 10 feet from where all of the connections come into the data center.  When I first joined Data102 the owners were fans of TMN, "We use it all the time!  It's our go-to test." and so they gave me a favorable deal and placement in their datacenter.  But it's still very expensive and I pay for that out of pocket.

 

Then I have about 30-60 VPS servers at any time, the majority of which right now are on Vultr VPS.  If you want $100 of credit to try their platform, here's the link.  I get $35 if you keep them.

 

I scale the VPS servers up and down with demand.  That's also expensive but I've paid much more in the past. 

 

I could definitely do it all for much less but the quality of the bandwidth is extremely important to me.  There are a lot of people who've come to rely on my service and I do my best to provide them with the best possible scenario from a server-side perspective.  I do extensive testing on hosts... xs1 just helped me throw out a promising prospect.  It looked fairly good but TestMy.net proved it's not as good as Vultr.  So when the client is fast enough, TMN can work in reverse and be a test of the host it's running on.  And I intend on doing an entire section where I test VPS hosts like this.

 

So many people blocking ads.  But oh well.

 

Quote

I now have test data from six of my computers, with up/down, multithread on/off, to three test servers, how do I summarize this to present to Spectrum? Spreadsheets are too complex.

 

When you have some computers pulling 250 Mbps in the linear test and others pulling 40 Mbps it's showing me that it might not be Spectrum at fault.  Are they delivering your 1000 Mbps... no, but I'm trying to tell you that 250 Mbps isn't all that bad.  In multithread it looks like you're able to see 400-600 Mbps.  Not bad.  If they're cool they should give you a credit for that.  Maybe $10-$20/month off for a year or something like that.  Comcast did that for me when I presented my information.  I just called and told them what I was finding, didn't have to show them anything.  They challenged nothing, but each ISP is different.

 

I've been a trusted resource going on decades now, your ISP knows TestMy.net.  They used to train their employees to use my site.

 

Quote

What are the reasons for such a wide variance between multithread on and multithread off?

 

• If it were happening on all of your computers across your network and also happening when directly connected to the modem.  Then it's probably a Spectrum issue.
• If it's happening, isolated to one or two machines... most likely a software or configuration issue of some kind.

 

Quote

Why do I have to turn off Privacy Badger in Firefox and Brave Shield in Brave to get your tests to work?

 

I test in Brave regularly, always with shields up.  You must have "block scripts" checked in the advanced settings or something.  If you block scripts here nothing will work right because all of the frontend code is javascript.  I don't know about Privacy Badger, probably the same kind of issue.

 

I'm getting back to writing my beta, where my time and effort is more appreciated.  Later!      << note: the thumb sticking out  

I'm sorry I just saw your topic.

 

40 Mbps is 5 MB/s which is 5120 kB/s

 

Mbps / 8 bits per byte = MB/s * 1024 = kB/s

 

So if you're trying to stream 8000 kB/s "you're gonna have a bad time"

 

 

For max quality I would set it at 4000 kB/s to leave headroom.  You should have no dropped frames.  You can technically go higher but I'd leave that 20% headroom for other stuff going on in the background.

 

Let me know if this works better.