Comments on: BattleGrounds Mod 2 / Rate caps in the source engine

By: raggi

raggi — Sat, 21 Apr 2007 16:50:57 +0000

Yes I do.

‘rate’ was made symmetric by valve quite a while ago (presumably, what in fact happened and was documented was cl_rate being removed), but the actual usage of client upload is significantly lower, and doesn’t spike in the same way as server data.

The ‘rate’ cvar is still documented in game as meaning the maximum rate the host can recieve data.

Furthermore it is not excessive upload that they require.

You can try and pick holes like that all you want, the fact is, it’s an unnecessary cap these days with regard to server to client data flow, and it needs to be lifted. Even 100 tick servers suffer issues with this value being this low. I can only presume this to be one of the reasons that Valve have not recommended 100 tick for a long time. I’m unsure if they’ve even ever retracted that.

By: GabeNewell

GabeNewell — Fri, 20 Apr 2007 16:54:35 +0000

Do you actually know what symmetric means in this context raggi?
The statistics/statements provided by Valve are completely correct, clueless :\

By: Dale

Dale — Thu, 08 Mar 2007 23:48:15 +0000

I might be wrong, but I think this is what rags is referring to ..

“The client creates user commands from sampling input devices with the same tick rate that the server is running with. A user command is basically a snapshot of the current keyboard and mouse state”

That’s taken from the wiki. there’s more on the source engine and networking, but i felt that was the most relevant part for what he referred to. Interesting read, if you’ve got the time for it.

By: raggi

raggi — Thu, 08 Mar 2007 14:14:45 +0000

Clients don’t upload physics. Clients upload controls and interaction information. Please see the Source engine documentation in the wiki before making more assumptions.

Furthermore net_graph, net_channels and rcon net_channels will give you some accurate numbers for this. Upload bandwidth for a single client is generally 50% or less of the speed of the average download rate.

By: h3iki

h3iki — Thu, 08 Mar 2007 03:25:05 +0000

untrue. would be true for HL1, however, the source engine physics give a lot more upload. still less, true. but not considerably.

i believe its upload that alfred is talking about :]

Admin note: Anyone, and I do mean ANYONE can simply load up the game and look at net_graph to see the difference between in and out bandwidth usage. This guys talking crap.

By: raggi

raggi — Tue, 06 Mar 2007 21:03:54 +0000

h3iki - the upload rate in source is almost always considerably less than the download rate for clients.

cl_rate should be re-introduced at this stage.

By: h3iki

h3iki — Tue, 06 Mar 2007 20:10:19 +0000

rate is both dl/upld. 240 kbps is regularly more than an average connection upload (eg. 512 adsl).

thus, while the dl wont be reached, the upload will be.

By: Dale

Dale — Tue, 06 Mar 2007 00:11:03 +0000

what was it you always say raggi? 40% of statistics are useless? ;p

By: Greg Hillenbrand

Greg Hillenbrand — Mon, 05 Mar 2007 23:12:23 +0000

Yah, I’m one of the Team Leaders from BG2. Thanks a bunch to Raggi, and, just in case anyone was wondering, Mr. Reynolds also said:

“Oh, and I should also add that sure, we will look at upping that limit (there are some fundamental limits due to memory buffering that will need to exist, but we should be able to make it a factor of 10 larger at a minimum I suspect). You will be shooting yourself in the foot if you use this larger number however.

- Alfred”

When I asked on a vague-ish timeframe, he said:

“When its done (within a month or so if I had to guess) and yup, we would post a changelog point about it.”