Optimization Coming in 1.0!

hey it's been a while remember me cuz I barely do uh anyways welcome back to yet another video it's been a while I missed you you know it's it's a long time since we since we got to hang out like this you know so uh uh you know I constantly get the question when we're going to release optimization updates and the truth is that almost every single patch we make contains some form of optimization it's sort of a never-ending battle but some changing have a bigger impact than others and today I'd like to share some of the work that's gone into improving the performance for the game coming in 1.0 before we dive into the topic at hand there are two things that I want to address first thing is that you may have heard news that our parent company embracer has sort of released a press release where they're saying they're planning to like transform into like sort of three Standalone publicly listed entities of which uh coffee stain and Friends uh working name uh uh the parent company for coffee s Studios is one of those uh I'm not going to go into detail about like this press release honestly because it's a little bit above above my pay grade uh you can find a link into the description if you want to read more on the on the press release uh if you're interested in that kind of stuff uh but there's been a bunch of speculation as to how this affects satisfactory and I just wanted to sort of come on here and sort of set the record straight on these things so so these changes to the company structure will not affect our plans of satisfactory at all we're still going to release the game this year year 2024 it hasn't changed the scope for 1.0 at all and we also plan to keep working on satisfactory after 1.0 is out whether that's in the form of DLC or you know free updates to the game is yet to be decided uh it might be either it might be both actually so so we'll sort of see but development for satisfactory is still going strong and actually speaking of which uh that's sort of the second thing I wanted to address here uh and it's sort of you know give you a little bit of a state of development because it's been a while since we published a video here on YouTube channel so um yeah we've been very busy behind the scenes managing the Clos beta for the select few that managed to get picked around a thousand is people have been playing the Clos beta now for a couple of weeks and the general feedback that we've gotten has been pretty positive actually and uh it's been a it's been really refreshing to hear all the nice things that people have said there's obviously a few technical things we all already knew that and there's going to be more of that down the line but so far everything's looking peachy actually so with this last bit of feedback we're now gearing up for the final push and you know hopefully it won't be long until we can maybe announce sort of a release de or something like that who knows exciting times anyways with that out of the way let's get to the main course and talk about optimization my favorite thing one big disclaimer is that performance is sort of tied directly to your hardware and configurations so some of these changes may be significantly more notable for some people rather than others so yeah just bear that in mind that performance is very varies a lot for from person to person uh also worth noting is that I'll put uh as much of this into layman's terms but there might be some technical jargon going on here and there as we go into details so just bear with me let's start talking about foliage my favorite subject everybody loves foliage so right now there are colliders for the vehicles which allow them to sort of detect contact with with any foliage and sort of destroy it as you're driving this results in pretty huge performances as like when you're driving through especially f dense areas but also in places where foliage is nearby but not necessarily in your path thanks to like wonky hitbox detection that we have uh in 1.0 most folage colliders are going to be completely removed from Vehicles actually uh which means that going forwards you can't destroy smaller foliage by driving it will simply just persist in the world uh which honestly also looks a bit better than when it just like vanishes into thin air when like the first pixel of your vehicle hits it so we we thought at first this was going to be a strange change but actually testing it out we felt like this kind of actually makes sense most noteworthy for dedicated servers folage has been like a huge culprit of Ram uh where sometimes it reached over like 12 gigs on like just a fresh safe file so not having Collision on foliage will reduce it down to like somewhere around like 5 and 1/2 gigs which is a huge Improvement relaxing folage Collision will also help greatly with level streaming and reduce overall stuttering because uh you're not trying to like demolish like dozens of objects within a second many times in a row all the time constantly so that's great but Vehicles will still be able to destroy almost all folage that requires at least the chainsaw to get rid of so like it's not gone totally but it's gone for like the majority of like the smaller folage types where the fine line goes between like exactly what folage is destructible or not is something that we're still sort of like figuring out and fine tuning so if you ever use belts then you'll be happy to know that we've optimized those significantly belts and pipes currently exist within like sort of unique instance groups in updated meaning that like each length of belts between connection points was sort of managed by themselves but in 1.0 they're going to be pulled together instead this is a big computational relief for your CPU as it won't have to work through as many instances of belts and pipes as it currently does so now instead belts and pipes uh belong to like fewer but bigger pools and now instead we're relying on the GPU to do the heavy lifting gpus can work extremely fast even when computations are chunky in size which is why for settled for this like fewer but bigger approach this helps stuff like Shadow rendering a lot since shadows will also belong to massive pools being calculated and drawn rather than like many individual objects or like smaller groups it's also great for instances where many small belts occur frequently which tends to happen a lot a prime example of this is manifolds where you know belts currently have to be computed per belt no matter how small they are uh which they tend to be in manifolds uh after this change again they're pulled together and the DPO will just like show it all through one thing that a lot of people ask us about is is whether we're going to use nanite for conveyor belts and we did play around a bit with nanite splines to try and like optimize the belt but in the end we sort of ended up settling on our own instancing solution because we found that it worked better for our game speaking of objects that may occur frequently everywhere we've now introduced a system that sort of converts several architectural buildables that players don't interact with uh or you know don't produce anything into Data only objects uh currently these buildables are what's referred to as heavy actors in a real like buildings that have like logic and actually do stuff uh the most common examples going data only would be you know foundations walls or architectural buildables like railings or stairs and such what changing the fancy term means is that we're basically cutting a lot of fat from a massive amount of stuff present in almost all factories so now we only store The crucial data for these things like locations the mesh that's using customization data such as like the colors and the recipe that we built with and the cool thing with all this like light data stored in a pre-prepared location in memory is that writing it to save data will be much faster so hopefully this will reduce the time the game stalls uh when saving when you have like a big old factory we're also skipping a bunch of steps when Gathering save data for these lightwe buildables uh related to production which also saves a bunch of uh time on that note uh lad times could also be a lot faster than 1.0 and and actually quite substantially faster uh because when all objects were heavy actors spawning them was costly because because we had a bunch of code that was like running through everything uh loading it all no matter if it was a foundation piece or if it was an assembler being spawned so having nonproduction build be lightweight allows us to skip a bunch of this code and you know set your files up faster and the actual safe files will also be a lot smaller non-production buildables being lightweight lightweight uh by the way I just realized I probably haven't pronounced lightweight properly but it's too late we're too deep into we got to keep going non- production buildable being lightweight also helps decrease runtime memory consumption you know actors are huge in unreal and with so many heavy actors turned lightweight uh enjoy not having like every inch of your RAM being you know used up by satisfactory this also connects to multiplayer because when actors are replicated to clients in Unreal Engine it requires that those clients allocate memory to track the state of the actors lightweight buildables means that this extra data is no longer needed so yeah tldr lightweight system makes stuff that don't have logic a lot smaller and easier to load and good and on the client side when playing multiplayer joining a game is also much faster now because processing those lightweights is so much faster than plowing through all these like heavy actors that existed before there's also a ton of General tweaks here and there that by themselves don't really amount to much but altogether add up stuff like reducing unnecessarily large texture sizes uh sorting stuff that was thrown together quickly for testing and then forgotten uh and fixing you know straight up broken in the world and and this means that vrm require IR ments are now a lot lower across the board scalability settings have also been looked at and adjusted so that people can expect you know a more reasonable result from their settings uh for example the performance hit going from medium to high in specifically sh Shadow quality settings was pretty weird so we've adjusted that and since the last update we've also upgraded Onre engine to 5.3 uh this will give the game a general sort of performance boost since you know epic has fixed a bunch of stuff on their end uh it also resolves a couple of crashes some people were running into and look okay I know what you're thinking okay 5 four just came out are you are you gonna upgrade to that one uh why aren't you upgrading uh well okay so there's a couple reasons why we're not going to upgrade to 5.4 uh it just isn't viable for 1.0 uh we might do it in the future but this is the version we're going to ship 1.0 with and for all you upscaler enjoyers out there you'll be happy to learn that we've updated dlss and FSR to their respective latest versions so if you're using those upscaling options you should also see a bit of an improvement there and finally in the last video we talked about new features coming to Dedicated servers well in the land of optimization dedicated service I've also seen some pretty massive improvements thanks to stuff like the foliage Collis thing and the pooling of belts and pipes however we're also experimenting with dedicated service to enable streaming up until now the way dedicated service has worked is that like it loads everything in the game so it loads the entire map and everything on it um and by enabling streaming for dedicated servers this means that everything on your server doesn't have to be loaded all the time so if you have like a bunch of people hanging out in like the same area uh you know loading that space should be enough until you start using the map more which obviously will sort of require much less memory when you do start exploring and building far away from each other uh then the stress of the server will sort of scale appropriately instead so those are some of the improvements coming in satisfactory 1.0 I'm sure there's going to be even further optimizations between now and final release but this is where we are right right now you may have noticed that there was a bit of a gap between this video and the last one you know there's like I said before there's a ton of stuff going on my scenes uh it just makes it really hard to push out these videos continuously right now uh but we will be back on track soon enough so please bear with us and until then I hope you have a lovely day stay optimized and take care yay optimization hooray oh I forgot to put on the beanie oh no did I record the entire video without the beanie oh no oh will people even recognize me oh man a bummer

Channel: Coffee Stain Studios

Views: 235,871

Keywords: satisfactory, satisfactorygame, game, games, videogame, videogames, indie, indie game, indie games, indiegames, factory, factories, windows, pc, openworld, singleplayer, multiplayer, coffeestain, coffeestainstudios, coffee stain, coffee stain studios, goatsimulator, goat simulator, sanctum, sanctum2, sanctum 2, fps, thq, thqnordic, thq nordic, epic, epic games, epic games store, embracer group, embracer, 1.0

Id: yIKjR5jVv9Y

Length: 11min 43sec (703 seconds)

Published: Fri May 03 2024