Page 4 of 4

Posted: 15 Mar 2007, 11:02
by Zydox
I still think the best way would be go get some kind of benchmark program into the spring program, so that it uses the same language as the rest of spring...

(I thought about having some kind of benchmark program for the lobby, but then I remembered that we have different lobbys... which I guess would yield different results on the same machine...)

I still think the main question is what's needed by the benchmark program... I know that the graphics card has some impact at the CPU usage...
As I've seen when I've been hosting battles with my AMD 4600+ with an integrated graphics card... It's always been about twice as high as my computer which was an AMD 3800+ (single core) with an NVIDIA 7900...

I would say that we need some help from the core developers to get us a benchmark program... if they're not up for it... skip it...

My personal favourite solution would be to be able to run spring.exe /benchmark and just get a value back... that the lobby then could use...

[song of the post: Icon Of Coil --- Repeat It (apoptygma berzerk remix)] :P

Posted: 16 Mar 2007, 17:30
by Licho
AF wrote:It would also be unreliable as it'd be highly dependant on the mod, map, and programs in the background. Even the players tactics could affect the results.
Well opposite is true for fixed benchmark too..

For example you can have player with computer of performance 1000 and a player with computer of performance 2000.

And player with weaker (1000) computer can still lag less than the other one.

Because it depends on stuff like actual resolution, actual settings, actual stuff on background when he plays, actual gaming style (how much zoomed is he, how often he scrolls).

Thats why I think that measuring these actual data from real games would give you better idea than fixed benchmark (like running some spring demo or using system info).

Posted: 16 Mar 2007, 17:38
by Licho
Zydox, it's in theory possible to make a benchmark SW that would use spring.

We could run spring with demo (using script.txt) and measure FPS using some hook on directdraw surface.

But .. huge drawback is, that it's not platform neutral.

But all what we need for spring benchmark is Spring that dumps average FPS into infolog on game end :)

Posted: 17 Mar 2007, 14:36
by monohouse
your lag problems are not because of low cpu or memory, it's because spring is badly coded, and im guessing it won't be better coded until all the bugs are fixed and all the necessary features are placed, until then there is nothing to do, getting better systems is only temporary solution for these versions, newer versions will get more laggy as more code is written, so if you really want to play spring without kicking players the only best solution is to wait until spring has a final version and only then see what should and can be done to fix it, as for osrts, that shitt is only gonna be worse as it is written in .net, so the only spring that will run fast enough on ANY system is the final version of spring, and of this spring not osrts

Posted: 17 Mar 2007, 15:53
by jcnossen
LOL
trolling is phail

Posted: 17 Mar 2007, 22:39
by AF
monohouse came here to argue? Its that poisonous people seminar on google all over again

Posted: 17 Mar 2007, 23:13
by iamacup
jcnossen wrote:LOL
trolling is phail
i swear i spread this word to the dev team via lordphail