Counter-Strike: Source

In CS:S, we chose the map de_train because it is fun to play around on. Source is a much more modern game than its predecessor, and as such it is CPU limited to a much lesser degree, but still has the limitation. This means that we should still see some sort of performance enhancement by decreasing the network CPU load.


Indeed, we did see some impressive results here as well. You can see that our average FPS increase was from 204 to 210, a 2.7% improvement. While the CS 1.6 increase was pretty insignificant due to the already insane frame rates, an increase like this in CS:S is far more tangible. The magnitude of this increase is still pretty small, but for gamers with mid-range to budget systems who struggle to get that elusive 100 FPS, the KillerNIC could do wonders.

Ping results for CS:S show improvements all around, albeit significantly less impressive than what we saw with CS 1.6.



You can see here that in Chicago servers, our ping decreased by exactly 2ms, an improvement of 3.4%. In Los Angeles servers, our ping decreased by only .08ms which, with pings already low, led to an improvement of a little over 1%. Strangely, this is the opposite trend to what we saw in CS 1.6. We believe we can attribute this to the netcode differences between CS 1.6 and CS:S, but again, we are far from networking or programming experts, so this is just speculation.