I generally map myself a little route through the map, then I'll do a base run (unmodified) using fraps' benchmarking tool (it gives you several CSV files with data), then I modify things for performance and try to stick to the exact same route again and run another benchmark. I can then quantify the differences.Ason wrote:So I tried this today, I stood in the center of the city, looking NE (which was the direction with lowest fps) I don't know how to get that red text in the corner of the screen that shows fps and stuff so I used fraps. I'm also not sure if this is how it should be done but this is the results:
with pavements: 10 fps looking NE, Average when walking around(center of map) 9-14 fps
without pavements: 18 fps looking NE, Average when walking around(center of map) 16-20 fps
Here's a very old example:
Another thing to consider is I think the lightmaps play into performance too, and so just deleting the statics will yield some performance gains, but you may get another couple of % gain when you re-lightmap the map without those statics. I'm not entirely sure, but it makes sense when looking at it from a drawcall perspective.
