Mar 15 2011, 2:58 pm
|
|
Lummox JR
   |
The server can't detect this issue because it's entirely client-side.
|
Lummox JR wrote:
The server can't detect this issue because it's entirely client-side. I know. That's why I asked how feasible it would be to make this information available to the server - is it worth making the feature request? It would be nice to be able to track CPU usage instead of always having to ask people what their CPU usage is. | |
world.cpu represents the amount of time spent processing DM code. The client doesn't run any(?), so what would client.cpu represent? The internal rendering and other engine related functions? Most of which we have no control over...
| |
It's not much different than something like this, which you seem to be on board with. There are certain aspects of BYOND that we don't have complete control over, but it would be nice to have measurements of their performance.
| |
If you're more interested in knowing what the general load is like on the clients, it's hard to say if that's feasible to calculate. I know how the server does this and I don't think there's really an analog on the client end. Obviously we could at least put in some kind of trip sensor so we could tell when the new keyboard-peek code was being executed, but that wouldn't be very accurate and might even trip in cases that would have been handled just fine in 480. Clients could perhaps report overall machine load (if that's even easy to get at in the Windows API), but that wouldn't be terribly useful to the server either because that depends on a lot of factors, and it's somewhat distorted by modern multiple-core machines.
I'm open to the idea of clients being able to periodically report their load--it might even give games the ability to do things like downgrade some effects for the end client, at least on a limited basis. The problem is getting CPU usage data that's reliable for a given process. I can always try to find out if that's something we can get or not. | |
Even of the number isn't very meaningful, it at least gives us something to measure. I can run the game, see what the client's average CPU usage is, make a change to the game, and see how CPU usage changes. Like measuring network traffic, the numbers might not mean a whole lot (its hard to define limits because they'd depend on other factors) but the numbers can be used to see if a change makes things better or worse.
| |