Unofficial UMK3 internet performance Thread

[nwo]

There has to be some point of diminishing returns doesn't there?

I don't think the average internet connection 10 years ago was anywhere close to what it is now, was it? And people played 06 UMK3 on Xbox Live and MAME and other internet games just fine. How does that and the fact that this game is 20 years old play as variables into it? It can't be as strenuous on a connection as MKX for example can it? There naturally just isn't near the amount of things going on, input wise, frame wise, background wise, engine wise, everything? Which would imply that you can play this game at an optimal level easier (on a weaker connection) then for example MKX? But I don't know?

The whole peering and bouncing concepts he described make a lot of sense. Very informative post, I didn't know most of that.

I don't know what the point of diminishing returns would be for UMK3. I feel like my connection is better than most players (not most of the guys who posted here), my upload speed is a little weak.

I don't know if MKX is more strenuous on the connection or not, just because the graphics are better shouldn't make a game more demanding on the internet. UMK3 is a pretty fast paced game also.

Was it just fine 10 years ago? Compared to playing offline I don't think it was ever really just fine to some people, and there seems to be a pretty big internet performance gap now a days.

 

[Ziggy]

 

[mkholic]

Are you trying to say though, that having ultra low internet speed doesn't make a difference and is acceptable?.

it's safe to say that with kaillera u can go as low as dial-up speeds and still surpass needed data exchage rete by far so it doesn't depend on the channel capacity at all

 

[nwo]

it's safe to say that with kaillera u can go as low as dial-up speeds and still surpass needed data exchage rete by far so it doesn't depend on the channel capacity at all

Dial-up? Ok.

 

[EdFig81]

Is Natas still alive ??

Lmfao yeah he is still alive . He recently had a kid and tbh hes not really into the fgc anymore.
His thing now is his kid, work and he is only playing destiny since it released . He hasnt touch mkx since a month or so after launch. We only played like 2 times and hes my roommate lol.
Hes mainly on twitter and facebook now.

 

[ClevelandMenace..JAX WINS]

I don't understand how to do this

 

[Drakonian]

 

[Konqrr]

Download and upload have nothing to do with how well your connection is online. It all has to do with your ping with the other player and quality of equipment from your modem to your system and theirs as well. Your network is only as good as the weakest point in it.

 

[nwo]

Download and upload have nothing to do with how well your connection is online. It all has to do with your ping with the other player and quality of equipment from your modem to your system and theirs as well. Your network is only as good as the weakest point in it.

I've thought about this for years, there's definitely a lot of factors. But, if you have higher quality more expensive net, a side effect usually is that your download and upload speed are faster too right?

I've witnessed the gameplay completely break VS people who have VERY slow internet, like under 1Mb/s upload. The few times I've had UMK3 desync in P2P have been with these same people.

 

[Konqrr]

I've thought about this for years, there's definitely a lot of factors. But, if you have higher quality more expensive net, a side effect usually is that your download and upload speed are faster too right?

I've witnessed the gameplay completely break VS people who have VERY slow internet, like under 1Mb/s upload. The few times I've had UMK3 desync in P2P have been with these same people.

Of course there is a limit to how slow the download/upload is to make the connection unplayable. I've had zero issues with playing online in several games with 1.5Mbps down for years (I have better now). I'm sure that their slow upload (< 1Mbps) has something to do with the desyncs. Someone with that slow of an upload should disconnect ALL other devices and phones from their network to give the game full attention, or enable QoS rules for the game to take priority.

 

[MRIGOTBASS]

Download and upload have nothing to do with how well your connection is online. It all has to do with your ping with the other player and quality of equipment from your modem to your system and theirs as well. Your network is only as good as the weakest point in it.

Going to have to disagree here professor.

http://www.calctool.org/CALC/prof/computing/transfer_time

We will use a typical DSL connection for example.

10 down / 2 up

The time it takes without ANY Latency in perfect conditions to transfer 1kb of data.

On the download side the connection can download 1kb in 0.762939ms <1ms
On the upload side the connection can upload 1kb in 7.62939ms almost 8ms

So you can see with a connection like DSL how the player with DSL can accept information 10x faster and see the moves the other player is inputting 10x faster than they can send the same information.

This is for a DSL connection lets look at a 56k modem rates now.

56K modem takes 17.4386ms to upload the same information with no latency and line conditions taken into the equation.

So yes its simple math, the connection 100% matters, and when its an asymmetrical connection its even worse. Those numbers are best case scenerio without taking into consideration any latency or other factrors.

Class dismissed again.

Edit: I am not sure what the packet size that Kalierra sends, i never watched it across the wire, however I can assume its smaller than 1KB for certain, however the frequecy of the packets per second can be as high as 60 depending on the network setting in mame.

 

[mkholic]

Going to have to disagree here professor.

http://www.calctool.org/CALC/prof/computing/transfer_time

We will use a typical DSL connection for example.

10 down / 2 up

The time it takes without ANY Latency in perfect conditions to transfer 1kb of data.

On the download side the connection can download 1kb in 0.762939ms <1ms
On the upload side the connection can upload 1kb in 7.62939ms almost 8ms

So you can see with a connection like DSL how the player with DSL can accept information 10x faster and see the moves the other player is inputting 10x faster than they can send the same information.

This is for a DSL connection lets look at a 56k modem rates now.

56K modem takes 17.4386ms to upload the same information with no latency and line conditions taken into the equation.

So yes its simple math, the connection 100% matters, and when its an asymmetrical connection its even worse. Those numbers are best case scenerio without taking into consideration any latency or other factrors.

Class dismissed again.

Edit: I am not sure what the packet size that Kalierra sends, i never watched it across the wire, however I can assume its smaller than 1KB for certain, however the frequecy of the packets per second can be as high as 60 depending on the network setting in mame.

u were on the right track, u see kaillera packets (having in mind it's on LAN setting) are less than 100 bytes on average so there's no problem in sending or recieving them at 60 pps
but hey, ur math efforts shouldn't be unrewarded so here u go

 

[MRIGOTBASS]

u were on the right track, u see kaillera packets (having in mind it's on LAN setting) are less than 100 bytes on average so there's no problem in sending or recieving them at 60 pps
but hey, ur math efforts shouldn't be unrewarded so here u go

Ok so if they were 100 bytes a KiloByte at least in the USA is 1000 Bytes so at 60 per second that would be 6000 bytes bka 6Kb. LOL

Not a pissing match and I believe you are dead wrong about the 100 byte thing.

Kaillera uses UDP and I believe a raw udp packet without a payload is 20bytes, so whatever the payload is add that on.

Download wireshark and run it capture the packets and tell us how big they are that way at least when you talk shit you will be armed with the real numbers.

My guess is around 50bytes per packet including UDP header information.

Point being is that it does matter, and based on your shit talking numbers it would definetly matter

Oh and I will crush you at UMK3 if you want to play lol

 

[mkholic]

Ok so if they were 100 bytes a KiloByte at least in the USA is 1000 Bytes so at 60 per second that would be 6000 bytes bka 6Kb. LOL

which is less than 56 kbit, lol
now u get how little it takes, funny, innit?

Download wireshark and run it capture the packets and tell us how big they are that way at least when you talk shit you will be armed with the real numbers.

last time i checked it was what i said it was, i didn't use wireshark iirc but some other packet sniffer

Oh and I will crush you at UMK3 if you want to play lol

i'd love to, i don't play with > 3 frames of delay tho and it won't be even that, u've got trikke to prove how good u're so chill

 

[UndefinedRican]

http://www.speedtest.net/my-result/4833361440
Don't know how you guys post the picture but yeah.