Any news on BT lagging issues ?

[Oracleman]

I have several players on my server still complaining about lag in BT and these persons never had any problems when we ran SH. Has anyone heard any rumors on when we can expect this to be fixed ?

Running Objective BTW.

[SNYPE]

Are they sure its lag and not just individual fps issues?
I am having trouble on my personal gaming machine with BT. My FPSs have dropped into the teens whereas with SH it never does that. For me I think its a problem with mboard, since tweaking the vid card has not helped.
Have them turn on the FPS indicator ingame and see what it does.
Bring up the console and type: "set fps 1"

[Oracleman]

I suggested that to them since I discovered I had to lower my own settings going from SH to BT. I'll return with any feedback.

Thx

[CFH666]

Last night we had enough people on to fully test our server. With 30 people on it, it ran nice and smooth, no one complained about lag, network or fps, however, when we played Berlin TOW, the pings average pings went from 70 to 140+, I was wondering why, so I remoted into our server and checked the CPU usage and it had skyrocketed during that map. Normaly with that many people it takes anywhere from 5 to 10% CPU usage, but when that map played it was consistant 25% usage.

Server runs on a Dual XEON 2.8 with a gig of RAM, and we still had roughly 65-70% of the CPU free, as well as almost 50% of the RAM free. I know that it is not do to networking, as ours is on a uncapped pipe and was only using approx the equivilant of two T1's. I know for a fact that our server has plenty of extra bandwidth that it shouldnt be network lag.

I know that the map does have a lot of scripting, with mortors and air strikes on it which Im sure adds quite a bit to the cpu usage, and it only did this on that map.

As for the other maps, I havn't noticed any lag or ping spikes on ours.



6

[XP-Tritech]

A lot of the lag is also being caused by the rate settings on the clients machines. If you did not set your connection speed properly you will be rated limited and will see lag.

XP-Tritech

[joe]

XP-Tritech

I run a BT server with 32 slots and 4 other SH servers (32 slots aswell) even though i force the rate to 8000-14000 on the clients its still lagging, on the same hardware with SH it runs whithout problems. I tried a bunch of settings but no success so far, any advice would be fine.

Hardware on the BT server:

Dual Athlon XP 2400, 1GB Ram (333Mhz)
20 Mbit up & Down.

When i look at the load on the CPU´s they use 45-50%, tried with default CFG settings and tried with custom settings.
My intension was to get another 2 BT Servers up and running but this makes me hesitate.

[sven]

XP-Tritech

The problem is not the network connection, the big problem is
the CPU usage, as soon as it start raise players get
big ping spikes.

[sven]

Reduced CPU usage would do wonder for this game.

[DV8R]

I dropped my display setting to 640 by 480 and texture to medium and its smooth as glass. Usually run 800 by 600 medium and high settings but Breakthrough doesnt like it with my 64 DDR card and the lower res actually doesnt look that bad if your advanced settings are high.


DV8R

[F_K]

HHmmmm well I have a whole different slant to throw in to the mix. I have been running a dedicated box from Artofwarcentral for some time now and it was running one 20 player Spearhead server on a Athlon XP2800+ with a gig of RAM. My processor ustlization was always between 30 and 60 percent with occasional spikes to 80 percent. I have people capped at 15k which I found to be optimal and with these settings ping was always really good and overall the server was running fine it was just an expensive way to do one 20 player server.

After switchiong to BT my processor utilization immediately dropped and it now runs a 20 player server with 15 to 30 percent utilization and I have never seen it rise above 40 percent. I have not run any older maps so there may be some issues there but for now it appears to be much better. I did however start to see peoples pings including myself shoot way up ( the server is in Atlanta and I went from an average of 30 ping to 80 average) and in heavy firefights there was some lag. This despite changing the network settings both up and down and I monitored this on another workstation while it was happening. To my surprise, althought the server would start to lag for everyone the processor utilization REMAINED the same. This was different than Spearhead in that there was always a direct correlation between server lag, ping spikes and processor utilization.
Puzzled by this I tried for the heck of it to raise the max rate cap on the server to 20k from 15k. In the past on Spearhead this would usually make the situation worse and 15k was the sweet spot but after raising it 2 days ago to 20k I can tell you the server now runs smooth as glass.
I don't know if this helps or make it more confusing but I thought I would pass on my experience.

BTW Tritech any word yet on the server commands????

[GaSplat]

Last night we had enough people on to fully test our server. With 30 people on it, it ran nice and smooth, no one complained about lag, network or fps, however, when we played Berlin TOW, the pings average pings went from 70 to 140+, I was wondering why, so I remoted into our server and checked the CPU usage and it had skyrocketed during that map. Normaly with that many people it takes anywhere from 5 to 10% CPU usage, but when that map played it was consistant 25% usage.

Server runs on a Dual XEON 2.8 with a gig of RAM, and we still had roughly 65-70% of the CPU free, as well as almost 50% of the RAM free. I know that it is not do to networking, as ours is on a uncapped pipe and was only using approx the equivilant of two T1's. I know for a fact that our server has plenty of extra bandwidth that it shouldnt be network lag.

I know that the map does have a lot of scripting, with mortors and air strikes on it which Im sure adds quite a bit to the cpu usage, and it only did this on that map.

As for the other maps, I havn't noticed any lag or ping spikes on ours.


I had a similar set up, and your problem is CPU usage. Yeah I know, its only 25%. Except for one iddy biddy detail. Xeons are usually run with hyperthreading - a trick that lets the OS see four CPUs instead of two. And you havve to change the monitor program options to show each CPU utilization individually, not just one total for CPU. Why? The game is not designed for SMP (multi-processor). So it uses only one of the "four" processors the Windows sees. That one is hitting 100% utilization and you are getting lag. Overall, three idle CPUs and one maxed out looks like 25% utilization, but most of that unused capacity is not available to your game server. If you have mulitple CPUs, you can add more games (up to one per CPU) with minimal performance impact. But they will all be smaller. I suggest you have the host turn off hyperthreading (BIOS setting and requires reboot). The OS will see only two CPUs - so your performance will improve for the one server, and you can still add a second with minimal impact.

[prisoner24601]

I had a similar set up, and your problem is CPU usage. Yeah I know, its only 25%. Except for one iddy biddy detail. Xeons are usually run with hyperthreading - a trick that lets the OS see four CPUs instead of two. And you havve to change the monitor program options to show each CPU utilization individually, not just one total for CPU. Why? The game is not designed for SMP (multi-processor). So it uses only one of the "four" processors the Windows sees. That one is hitting 100% utilization and you are getting lag. Overall, three idle CPUs and one maxed out looks like 25% utilization, but most of that unused capacity is not available to your game server. If you have mulitple CPUs, you can add more games (up to one per CPU) with minimal performance impact. But they will all be smaller. I suggest you have the host turn off hyperthreading (BIOS setting and requires reboot). The OS will see only two CPUs - so your performance will improve for the one server, and you can still add a second with minimal impact.

It seems strange that inorder to improve server performance you would have to disable an advance feature (HyperThreading or HT). Although BT may not be coded for SMP, simple facets of the Operating System are. Even if you disabled HT wouldnt the same load be placed on CPUs still? Instead of showing 25% CPU usage for 1 out of 4 CPUS, you get 50% now for 1 out of 2. The ratio is still teh same. And wouldnt it be crippled somewhat cuz HT is disabled?

[GaSplat]

The difference is one CPU in HT is a quarter of the total system capacity and one CPU without it is half the total capacity. Yes hyperthreading is an advanced feature - but it is designed for systems that have a lot of little tasks running at once. It reduces the frequency with which a system has to "change context" for a CPU as it bounces between the different tasks. Changing the context involves a fair amount of overhead for the operating system.

In the case of a game server, we are not dealing with lots of little tasks that run independantly in their own context, but one CPU cycle hungry beastie that just wants more, more, more. While today's CPUs can handle the majority of little tasks "one handed", Spearhead and Breakthrough demand its full attention. That's because all the "threads" and entities the game spawns are all playing in the shared context of the game environment, and the OS cannot break them out into separate tasks to distribute amongst CPUs. Instead the game engine handles that, acting as one big task for the operating system, who gives it to one CPU. Make that CPU be a single physical CPU that is just doing that one task, and you have an edge over a physical CPU that is only offering up two "half strength" virtual CPUs to the operating system (that is the simple explanation of what hyperthreading does).