I said this on Masto, but this tells me nothing as written. You can get the first game to run like that, too.
The thing is, if it runs that way on an empty map and degrades the same way the first one did, I can't see it not crashing on a full endgame map. So... how does it run on endgame? Or is this endgame and it runs fine at first? Guessing no, since the devs themselves said this was a problem. And, well, I've seen footage from streamers and it certainly chugs on small maps, too.
To answer your main Q from this, apparently the biggest performance dropoff is going from no pop to 10k, scaling up to 100k isn't nearly the same dropoff.
Even with a beefy setup on high settings, CPP suggests turning off many of the post-processing effects and definitely disable VSync. Lastly 4K is out of the question for most cards, barely playable for top end enthusiast cards, Most will be limited to 1080p for a usable experience.
The good news from this video is that for anything above a 970 it is possible to get to late game, as long as you stick to 1080p low-med settings on any card with less than 8GB VRAM.
Hm. So it scales with VRAM and GPU, not CPU? Interesting.
That's less concerning than people had made it out to be, at least for a game of this genre. It still doesn't sound particularly pleasant to play, but hey, less of a dealbreaker.
Not that I don't believe them, but it's odd that none of the videos from YouTubers with early access have shown that kind of performance. It makes me wonder if they are trying to set lower expectations for some reason.
I was incredibly excited for this game because I thought it'd be an actual city simulator this time, but it's just another American-road-based traffic simulator.
It's not going to be worse than KSP2. KSP2 launched with game breaking bugs and all of the new features missing (and some of the old ones) as well as bad performance on ALL graphic settings no matter the specs. CS2 has features that differ it from the first game and the performance issues (allegedly) can be fixed by turning settings down (unlike KSP2). Both games launched (or will launch) without modding support which is really bad for both games.
I feel like most people above 1920x1080 actually rock a 1440p setup cause it's a serious step above traditional HD but without the needs of a 4k capable GPU.
There is actually a LOT more little things, that make the game very different. Besides who in the right mind would buy a 70 buck sequel with Just better grafics?
Just have a look at the dev diaries. For me personally its the overhaul of pretty much all simulation engines (traffic, weather, water, wind, people etc.) and that they solved (apparently) the single thread problem of their traffic simulation. For me CS1 was bottlenecked when the cities became to big and the traffic could only be simulated on one core. There is a limit to that. But my cpu was otherwise idle. I have hope that this is now solved. Plus there is apparently no agent limit anymore. So a town of 500000 could in theory simulate all people individually, CS1 couldn't.
I hope Digital Foundry does a review. I wanna see CPU utilization so badly, Paradox needs to learn to invest into CPU optimizations for their CPU heavy games
They do invest a lot in cpu optimization. The problem here seems to be unoptimized GPU performance.
In addition, you will always struggle with CPU performance in complex simulation games with many interlocking systems. There's only so much you can do without limiting the gameplay.
Considering skylines is basically the only surviving city sim franchise, not much. But city sims have always had difficulty with performance. Sim city 4 was notorious for how badly it performed in hardware, even to this day