Rimworld is one of the few games for which DLC is an insta-buy for me; every one of them has been a pretty fantastic expansion. This one appears to be adding on quite a few interesting mechanics -...
Rimworld is one of the few games for which DLC is an insta-buy for me; every one of them has been a pretty fantastic expansion. This one appears to be adding on quite a few interesting mechanics - beyond the included content, I'm excited to see what the modders do with the new tools once they've had a chance to play. The inclusion of books in particular seems like it will enable some truly cool research and knowledge mods.
I mean there are some horror mods that can use the new mechanics that Anomaly introduces, something like what happen with the race mods integrated Biotech in their mods. So, I am looking forward...
I mean there are some horror mods that can use the new mechanics that Anomaly introduces, something like what happen with the race mods integrated Biotech in their mods. So, I am looking forward to the new ending and mechanics that it introduces. Along with the new mods that people will be making.
Dang, that looks exciting. Rimworld was my pandemic-unemployed-and-depression game, I put over 800 hours into it over the course of a year or so and I'll always remember it fondly, but it's...
Dang, that looks exciting. Rimworld was my pandemic-unemployed-and-depression game, I put over 800 hours into it over the course of a year or so and I'll always remember it fondly, but it's (thankfully) hard for me to recapture those conditions. It's one of the most interesting and flexible colony sims I've ever played, and the modding community takes it to real greatness. If they ever added proper multiplayer I may just play it forever
Having tried playing it a couple of times I can confidently say Rimworld is one of those games I should never own, because I can easily see myself starting games and then sinking far more time...
Having tried playing it a couple of times I can confidently say Rimworld is one of those games I should never own, because I can easily see myself starting games and then sinking far more time than is healthy into.
Civ is like this too. Just one more turn... and then 6 hours later I've given myself a headache and eye strain. My only real solution has been to just not play.
Basically all of the stuff that's not DLC specific is them mainlining popular mods, but I'm not mad. It's worth them polishing up the best ideas and promising that they'll work out of the box with...
Basically all of the stuff that's not DLC specific is them mainlining popular mods, but I'm not mad. It's worth them polishing up the best ideas and promising that they'll work out of the box with no depending on modders or a delay for mods to update after release.
Tempting, because they're finally multithreading it. The monster part isn't so much appealing, but all the other bits in the 1.5 update are. Colonies used to bog down, a lot about the time they...
Tempting, because they're finally multithreading it. The monster part isn't so much appealing, but all the other bits in the 1.5 update are. Colonies used to bog down, a lot about the time they started to get to the point where I felt they were semi self-stable.
The way I played, that was usually about twenty-thirty colonists, to cover all the jobs. I always played it as a colony simulator, not as a tactical thing, so I was always loading in mods for farming and cooking, and I'd have fun giving everyone their own room, and a variety of meal and pleasure options, and all that. Just building up a nice stable self-sustaining outpost.
Tons of fun ... until ... it ... chugs ... to ... a ... halt.
Which it always did. I tried everything, but there's only so much you can do when it's ignoring all the other cores on the CPU. Every management game has this problem. Why don't programmers multithread? It's not like it's a new thing; consumer level multi-core CPUs have been around for decades now. And even Rimworld, even now, took eleven years to get around to multithreading.
Still, better late than never I guess.
Nice to see the Warcrimes Simulator is still chugging. Guess over the next few months I'll catch up on some Let's Plays where they combine the monsters with all the usual suspects (organ harvesting, prisoner torture, you know, the stuff Jack Bauer would approve of.)
They often do, but the problem is that most things in games cannot simply be easily multithreaded. The Factorio devs have written a ton about their efforts to improve performance, and address the...
Why don't programmers multithread?
They often do, but the problem is that most things in games cannot simply be easily multithreaded.
The Factorio devs have written a ton about their efforts to improve performance, and address the question about multithreading really well. The biggest problem is that stuff that's deterministic generally can't be multithreaded because it all relies on other computations being completed.
And, not to burst your bubble, but they've only multithreaded certain types of drawing, so I doubt you're going to see major improvements in late game colonies with a couple dozen pawns. I really doubt that's where the main bottleneck is in late game colonies.
Random object access is rarely threadsafe. Gameplay is easy to make spectacularly single-threaded. By the time you realize that your gameplay pipeline is the bottleneck of the app, it's so...
Why don't programmers multithread?
Random object access is rarely threadsafe. Gameplay is easy to make spectacularly single-threaded. By the time you realize that your gameplay pipeline is the bottleneck of the app, it's so morbidly attached to the main thread that pulling it all apart and designing it to cross threads is a nightmare that the budget can't sustain.
And for the followup, why don't game programmers design their gameplay logic for thread safety from the get-go, I'll again point to tiny budgets. Game devs run as fast as they can to get a working prototype that proves the gameplay so they can sell the project to a publisher. The publisher then says, "Cool game, here's not-enough-money to make it saleable." Now the developers have only enough cash to hire bargain bin engineers who will be swamped just polishing up the prototype and supporting the inevitable feature-creeps that the publishers ask for. The idea of revamping systems for proper simd is not feasible.
Why yes, I am a bit bitter, having just gone through this exact scenario.
This is why game engines need to support simd with their default tools. Cash-strapped devs can't afford it. I was super interested when Unity announced HPCS/DOTS but they seemed to have stopped it. Unreal has some good threaded-by-default stuff for rendering and audio, but most gameplay takes place on the main thread still.
Ludeon is in a unique position, though, since they likely have tons of cash from the success of Rimworld and its DLCs.
Simulation games like this typically struggle to multithread the performance heavy parts of the game because they rely on other information that needs to be computed on the same thread or...
Simulation games like this typically struggle to multithread the performance heavy parts of the game because they rely on other information that needs to be computed on the same thread or otherwise needs to be computed in ways that make multithreading difficult (I'm not an expert on the actual mechanics here, only that it's a common problem in this genre in particular). Dwarf Fortress and Oxygen Not Included, the two other colony sims I play, have the same or similar problems.
Having fewer colonists is often an easy way to avoid the worst of this, since computing things related to their states and actions is often a pretty big computational burden. Rimworld actually does better at this than ONI or DF, since the Storytellers target a specific range of colony sizes and will tweak things to compensate. I know that in both ONI and Rimworld lag can be reduced by limiting animal pathfinding as well -- either through keeping numbers low through regular slaughter or confining large groups to a tiny area to reduce the computational load of pathfinding for them. Rimworld also has the advantage of not trying to simulate fluids, unlike the other two games, where fluid simulation can also contribute to lag in certain extreme cases. I think temperature simulation is a contributor to lag in Dwarf Fortress sometimes; I wonder how much it impacts Rimworld in that way.
Rimworld is one of the few games for which DLC is an insta-buy for me; every one of them has been a pretty fantastic expansion. This one appears to be adding on quite a few interesting mechanics - beyond the included content, I'm excited to see what the modders do with the new tools once they've had a chance to play. The inclusion of books in particular seems like it will enable some truly cool research and knowledge mods.
The modders having more tools is absolutely my favorite part, but I do love horror and am glad to see it.
I mean there are some horror mods that can use the new mechanics that Anomaly introduces, something like what happen with the race mods integrated Biotech in their mods. So, I am looking forward to the new ending and mechanics that it introduces. Along with the new mods that people will be making.
Dang, that looks exciting. Rimworld was my pandemic-unemployed-and-depression game, I put over 800 hours into it over the course of a year or so and I'll always remember it fondly, but it's (thankfully) hard for me to recapture those conditions. It's one of the most interesting and flexible colony sims I've ever played, and the modding community takes it to real greatness. If they ever added proper multiplayer I may just play it forever
Having tried playing it a couple of times I can confidently say Rimworld is one of those games I should never own, because I can easily see myself starting games and then sinking far more time than is healthy into.
Civ is like this too. Just one more turn... and then 6 hours later I've given myself a headache and eye strain. My only real solution has been to just not play.
Basically all of the stuff that's not DLC specific is them mainlining popular mods, but I'm not mad. It's worth them polishing up the best ideas and promising that they'll work out of the box with no depending on modders or a delay for mods to update after release.
Tempting, because they're finally multithreading it. The monster part isn't so much appealing, but all the other bits in the 1.5 update are. Colonies used to bog down, a lot about the time they started to get to the point where I felt they were semi self-stable.
The way I played, that was usually about twenty-thirty colonists, to cover all the jobs. I always played it as a colony simulator, not as a tactical thing, so I was always loading in mods for farming and cooking, and I'd have fun giving everyone their own room, and a variety of meal and pleasure options, and all that. Just building up a nice stable self-sustaining outpost.
Tons of fun ... until ... it ... chugs ... to ... a ... halt.
Which it always did. I tried everything, but there's only so much you can do when it's ignoring all the other cores on the CPU. Every management game has this problem. Why don't programmers multithread? It's not like it's a new thing; consumer level multi-core CPUs have been around for decades now. And even Rimworld, even now, took eleven years to get around to multithreading.
Still, better late than never I guess.
Nice to see the Warcrimes Simulator is still chugging. Guess over the next few months I'll catch up on some Let's Plays where they combine the monsters with all the usual suspects (organ harvesting, prisoner torture, you know, the stuff Jack Bauer would approve of.)
They often do, but the problem is that most things in games cannot simply be easily multithreaded.
The Factorio devs have written a ton about their efforts to improve performance, and address the question about multithreading really well. The biggest problem is that stuff that's deterministic generally can't be multithreaded because it all relies on other computations being completed.
And, not to burst your bubble, but they've only multithreaded certain types of drawing, so I doubt you're going to see major improvements in late game colonies with a couple dozen pawns. I really doubt that's where the main bottleneck is in late game colonies.
Random object access is rarely threadsafe. Gameplay is easy to make spectacularly single-threaded. By the time you realize that your gameplay pipeline is the bottleneck of the app, it's so morbidly attached to the main thread that pulling it all apart and designing it to cross threads is a nightmare that the budget can't sustain.
And for the followup, why don't game programmers design their gameplay logic for thread safety from the get-go, I'll again point to tiny budgets. Game devs run as fast as they can to get a working prototype that proves the gameplay so they can sell the project to a publisher. The publisher then says, "Cool game, here's not-enough-money to make it saleable." Now the developers have only enough cash to hire bargain bin engineers who will be swamped just polishing up the prototype and supporting the inevitable feature-creeps that the publishers ask for. The idea of revamping systems for proper simd is not feasible.
Why yes, I am a bit bitter, having just gone through this exact scenario.
This is why game engines need to support simd with their default tools. Cash-strapped devs can't afford it. I was super interested when Unity announced HPCS/DOTS but they seemed to have stopped it. Unreal has some good threaded-by-default stuff for rendering and audio, but most gameplay takes place on the main thread still.
Ludeon is in a unique position, though, since they likely have tons of cash from the success of Rimworld and its DLCs.
Simulation games like this typically struggle to multithread the performance heavy parts of the game because they rely on other information that needs to be computed on the same thread or otherwise needs to be computed in ways that make multithreading difficult (I'm not an expert on the actual mechanics here, only that it's a common problem in this genre in particular). Dwarf Fortress and Oxygen Not Included, the two other colony sims I play, have the same or similar problems.
Having fewer colonists is often an easy way to avoid the worst of this, since computing things related to their states and actions is often a pretty big computational burden. Rimworld actually does better at this than ONI or DF, since the Storytellers target a specific range of colony sizes and will tweak things to compensate. I know that in both ONI and Rimworld lag can be reduced by limiting animal pathfinding as well -- either through keeping numbers low through regular slaughter or confining large groups to a tiny area to reduce the computational load of pathfinding for them. Rimworld also has the advantage of not trying to simulate fluids, unlike the other two games, where fluid simulation can also contribute to lag in certain extreme cases. I think temperature simulation is a contributor to lag in Dwarf Fortress sometimes; I wonder how much it impacts Rimworld in that way.