A bit of time and testing has passed since my last post. I admit, I’d not actually intended for this beta to progress beyond more than a simple experiment, but with the event of the 500+ unit map test exposing lag and stutter oddly and the chance to try out my vehicle proxy class idea we actually have what I hope is a significant improvement.

In the above video I have what I hope is convicing evidence, albeit not totally scientific, of the progress I’ve made in optimizing Revolution.

The test map in this case is probably a bit unusual for most players but perhaps not so for someone really getting into the single-player PvE building their own little empire. There are around 200 player owned units, mostly land turrets. There are a few barges and aircraft and one carrier. The other 300+ units that are present are the AI controlled units on nearby islands.

My system is on paper basically the min spec for CC2. I have an i5-2500 which is actually slightly higher clock rate than the i5-4440, my GPU is a trusty Nvidia GTX-970 which is above the recommended spec.

Constrat reported he had a far superior system, a modern i7-6800k and a Nvidia RTX 4070ti so I was quite surprised at the time that there was any local lag at all.

Now that I’ve I think squeezed a good deal of performance out of my lua code it’s worth looking at how the game feels when really under stress all in the same scenario but without mods, and with 3 different versions of Revolution:

My test is simple

• Load the map
• Run around the bridge once
• Enter a control screen
• Pan/zoom to see all the map
• Stop

Just running this simple loop in all 4 scenarios shows the differences doing extra modding can make to the bridge screens.

I think you’ll agree that Vanilla is clearly a lot faster and smoother, but that 1.5beta2 gets 2nd place, possibly only because of my mouse twitch at the start.

—

Bred