Jump to content

faatal

Fun Pimps Staff
  • Posts

    2,947
  • Joined

  • Last visited

  • Days Won

    104

Posts posted by faatal

  1. 2 hours ago, khzmusik said:

     

    Any chance to allow XML configuration on the next pass, whenever that is? Even if it's just a flag to not display the tool?

     

    I know a lot of modders like to add (or revive) different tools for workstations, and this would help with immersion.

    You can now do this which I'm putting on the forge xml as a hint (is default without it defined).

    <property name="ToolNames" value="1,2,3"/>

  2. 2 hours ago, Guppycur said:

    What's the xml look like on that, i.e. will we be able to target some lines and use our own meshes?

    No xml. The prefabs are simply setup to have a child object named 1 for for tool slot 1, named 2 for second and so on.

  3. Friday, I started on the forge mesh showing/hiding each tool and it seems to be working. This will probably make it into A21. Code supports all workstation types, so will be simple enough to do the rest as new models/features are added.

  4. On 4/15/2023 at 5:53 PM, Riamus said:

    Yeah, I can't imagine they do it except between traders.  I wonder how they would manage vehicles... Do you have to pick it up to take it with you, or having it on the helipad is enough?  We shall see.  Since they did remove it, that means it is probably guaranteed to appear.  No reason to remove it otherwise.  Of course, maybe they just don't want to have false hope out there. 😁

    It should be no surprise to anyone that we discuss fast travel as many games have it, hence the helipads. The vehicle teleport is one of my main issues. We think the vehicle should go with you without it being annoying. I'm thinking it teleports to an open parking like area and finding a free spot. I'd prefer it be a very limited system, but it could end up as whatever makes sense or not doing it at all. It could easily not be in A22, yet show up later.

  5. 13 hours ago, meganoth said:

    Just an idea: One way he could achieve this would be moving calculations to the clients. For example **IF** zombie pathing currently is all done on the server, in his mod each client might do the pathing for the zombies summoned by that clients player.

    Zombie pathing is already cheap and runs off the main thread. I did fix a bug last week with only 1 path per frame being sent to the threading system and since it typically only takes 1 frame to make a path, it was a bottleneck for say 100 zeds making paths each second. A fairly simple change now has it requesting up to 8 paths per frame of which the threading system can provide about 6. This tested well with around 150 zeds on a server spread across 6 players.

  6. 13 hours ago, Blake_ said:

    @faatal I have a couple of questions:

     

    1.How many options does the Chunk reset option have? I mean I'm curious if there are just simple ones like 10,25,20,50 days (and disabled) or if more advanced ones like "enable only on chunks without player-made blocks (resets if all player-made blocks get destroyed" or things like "do not reset land claim areas", etc, are a thing.

     

    2. How better does 7dtd perform in a21 vs a20? did you noticed significant gains near trees, water, Zombies and/or Blood Moons in general? 

     

    Thank you in advance. 

    1 It is just a reset time in days. It works at the POI/chunk/region level. The engine does not track player made blocks.

     

    2 Hard to say. So many things change that you can't just load an old world and stare in the same direction and see if it is x FPS faster, because that world won't load and if it did, POIs have changed or new blocks and some blocks converted to props. A lot of it is situational with where you are, what you are doing, your hardware and your options. The simple answer is it runs a bit faster and RAM use is a fair bit less.

  7. 4 hours ago, Roland said:

    There is a difference between working steadily on bandits since A15 until today and starting on bandits for A15 and then putting them aside until the start of A21 work. The first statement denotes about 5 years of working on bandits while the second more accurately puts it at about a year and a half.

    Yes.

     

    I started at TFP at the beginning of A17 and later in that alpha, took over AI and made large changes to how AI for zombies and animals worked. Until A21, there were dozens of big things and hundreds of smaller things for me that were higher priority than bandit AI. The new bandit work, which started during A21, is based on the current state of the AI systems, not prototype code/data from 6-7 years ago.

  8. 18 hours ago, HeLLKnight said:

    In case @Faatal you say again no to HDRP, some people are about to release in a few days a functional FSR 2.2 support working in Classic/Built-in pipeline to Unity Asset Store. This would be a big win for all of us.

    https://forum.unity.com/threads/boost-your-fps-with-amd-fidelityfx-super-resolution-2-0-fsr-2-support.1357190/

    Interesting. I'll keep an eye on that thread. If it works well, then we may support it.

    Thanks for the info.

  9. On 4/14/2023 at 5:51 PM, Darthjake said:

     

    Yeah, but you note he didn't say how big the map was, I mean 4k maps can be ready in 5 min or so depending on how many cities you've got programmed in.

     

    I made an 8k map the other day and it took 3.5 mins. Granted, this is on my new PC, which is a AMD 7950x.

  10. 50 minutes ago, Riamus said:

    Does that mean you still cannot full a pool or similar enclosed space with water in game more than about 1/4m high?  That is unfortunate.

    No. That is the water surfaces level in a voxel, not how many voxels deep. We have bodies of water of a variety of depths. The water mesh is only created for the top, otherwise you would have layers upon layers of meshes, leading to visual issues and slow rendering.

  11. 2 hours ago, Cr0wst0rm said:

    Like the single blocks that very sticking out... I would expect it would flatten the center and fill surroundings... 10% height of a block? Almost I am feeling like if the solution that Niagara uses in Unreal Engine 5 (volumetic simulated water) would be better in conjunction with the voxel collision stuff... But performance issues would probably come with that....

    They showed the worse use case. The current water is meant to be contained, not dumped on flat ground. Neither RWG or POIs have water dumped on flat ground.

    The water mesh generator is the old mesh code with some cleanup and the programmer working on it is still making changes. The mesh generator makes meshes at a fixed height, which does not tapper off currently.

    18 minutes ago, Survager said:

    @faatal Can we get answers to these questions from you? if you know what to answer, of course )

     

    1. Will the 20.7 update that was announced on the stream break old saves made in version 20.6?

    2. Will players need to create a new game after Alpha 21 is released if they start a new game after 20.7 is released?

    1 - A20.7 has a few minor changes to reading the world file header. It will not effect existing A20 saves. It will simply show A21 saves in the list without throwing errors.

     

    2 - A20.7 saves are the same as 20.6 saves. A21 will not read those saves due to many changes that are not compatible.

  12. 11 hours ago, Riamus said:

    They also said A20.7 will come out Thursday, which I guess is later today now, to fix a bug that will occur when swapping between experimental and A20.  Can we get verification if updating to A20.7 will affect saves?  It sounds like it only is there for that one fix and so *shouldn't* affect saves, but I want to check.

    A20.7 has a few minor changes to reading the world file header. It will not effect existing saves.

    6 hours ago, DocRussel said:

    My favorite part was Rick going against the others' wishes and actually giving a rough estimate of early-mid May for the A21 experimental release.  Of course that can change, I'm just thankful for even an educated guess on their part as of right now.

    As I said a few times in the stream's chat, it will be x amount of weeks. Rick gave a more specific guess, but really it could be 3 weeks or 5 weeks or more and comes down to fixing any bugs that we do not want in even an experimental release once we finish features and polish we are doing right now.

     

    There are 63 must fix tasks right now, of which maybe half need to be done to release experimental and more will be added as we go.

  13. 4 hours ago, Blake_ said:

    Well, zooming and grey area cleaning are both done by the GPU I gather, but the map data isn't, so long story short, there just shouldn't be a technical limitation anymore, it's just that they didn't touch zooming since ages ago. The map "blacking out" when exploring lots of it? I haven't tested it yet in this alpha.

     

    Speaking of which, @faatal do you know if there are changes regarding map exploring/ zooming in a21? 

    Also, Can we fully explore a 10k or even a 20k custom map in multiplayer and will it remain explored even after thousands of hours of restarting the session and spitting at our motherboards?

    It is the same. The limit is it renders into a texture and the texture is a fixed size to not use too much memory.

     

    I did change zoom a while back in Unity editor, so I can teleport around better or zoom in very close to check the art, but it is not for release as the texture no longer fills the window, so you would see the edges. No plans to change.

    Never looked at how it saves/loads the explored area.

    3 hours ago, Annihilatorza said:

    @faatal Do you use Frustum culling?

    Every 3d game engine uses culling by the camera's view. Our occlusion manager goes beyond that and culls based on areas being hidden behind rendered objects.

    5 minutes ago, Trankitas said:

    I hate to bring bad news, but every Alpha until A20 needed a fresh start, but imagine all the new props and generated cities! It's a good thing if you think about it, you can save your world by backing up %appdata%/7daystodie folder just in case, if you want to play it again just go back to A20 (properties/betas) on steam.

    Yes, A21 has many changes that would break an A20 world.

  14. On 4/1/2023 at 8:45 PM, Roland said:

    None. Just a couple of chatAI programs they left running when they left the building.

    According to ChatGPT we are released.
    "...as of my knowledge cutoff of September 2021, 7 Days to Die is already a released game. The game was initially released in 2013 as an early access title and has since had several updates and full releases."

  15. On 4/1/2023 at 10:47 AM, Jost Amman said:

    @faatal I know it's Saturday, so I won't be surprised if you can (eventually) reply only next week... but I was curious to know if there's any open ticket to fix the bug where hatches don't stop rain. I've just now read a post on Steam, of someone complaining about this, and that this bug has been around forever. I agree this is long due IMO.

     

    Is there any hope to put this in the bucket list for A21?

    Thanks in advance :)  

    It may already be fixed in A21, because I made changes to the rain particle system physics settings to stop it going through roofs.

  16. 2 hours ago, adny said:

    I wanted to convey to you, one interesting thought!
    It seems to me that your concept of developing equipment, workbenches, furnaces, weapons is, in principle, not true!
    The fact is that even I can’t make the machine itself on my knee. To do this, you need at least 2 machines, turning and milling. Nobody! No one can make these machines! No one can make an automatic machine, repair it, yes, make it from scratch, no.
    Therefore, it is necessary not to do all this, especially after the apocalypse, there will be so much of this goodness ....
    It is necessary that the player finds, machines, furnaces and so on. He brought all this to the shelter and repaired it, restored it. You can somehow learn this, do it from scratch - NEVER!
    By the way, this applies to cartridges, no one can make them. Yes, according to the recipe in the game, they will be made, but they will not shoot!
    And why???

    It is a game and the number one rule of professional games, is to make it fun, so people will buy it and play it enough to tell their friends and repeat the cycle or you won't be making games for long because you will go broke. Most games, including 7dtd, are far from a real simulation of anything, and I'm OK with that. Where we can make it more realistic, sure, we try, but if it takes too much dev time or ruins the fun, or is overly complicated where many players would be turned off, then we skip it.

    2 hours ago, Rabbitslovecactus said:

    It would be really cool if when you cross the map boundaries of Navezgane, it goes into RWG. That way you can enjoy the best of both worlds. Also have an option to enable and disable that feature. 

    That would require some serious engine changes, short of we just load the RWG map and drop you into it.

  17. 22 hours ago, Mr. Perfect said:

     

    Sorry to bring up old topics again, but the new DirectStorage 1.1 release might solve that processing speed issue for you. Anandtech did a great little article on the 1.1 version, but here's the bits that would probably interest you.

     

     

    Like you said, Unity would have to do the work of adding DirectStorage to the engine, but it sounds like something 7 Days would benefit from. Maybe give your Unity rep a nudge?

    The processing I was referring to is our c# code converting data from say xml into bytes in memory, which a GPU is not going to help with. Reading of xml also causes a lot of secondary data to be read like loading of prefabs, which loads components, meshes, materials, shaders and textures. All of that being done by Unity.

     

    Asking Unity for features generally have 6 months to years of lag time before you would see it in an LTS, which is the only versions we use these days.

    20 hours ago, falloutcloud said:

    I think the reason many of us had hoped it would release by Christmas is that Rick sounded optimistic that it could happen. We all pretty much knew that it was just a guess but I don't think that we thought he'd be off by 4 months.

    Well I was optimistic too, but all it takes is one dev working on a core required feature who thinks they will be done in a month and then unforeseen problems or design changes happen and it becomes 4 or 6 months.

  18. 2 hours ago, Abunai said:

    From what I understand from your comment and previous reports @faatal, there are two serious bugs that need to be fixed for the release of A21:
    This zombie AI error you mentioned and the random hole in the ground of POI when quest resetting. Please, can you share with us something about the status of this second bug? (sorry if that was answered previously)
    Thank you and good work 

    Those are only my bugs and not all of my bugs/tasks as I currently have 12 must fixes in various states of completion and across the team there are 67. That number does not mean a lot, as half of them do not need to be done for experimental and new ones get added regularly.

     

    Then there are last minute tweaks, like today it came up that sleeper volumes spawn using the game stage of the player triggering the volume, so I am trying to test a change that adds in players nearby, similar to how blood moons spawning makes parties from nearby players for game stage calc.

  19. 8 hours ago, beHypE said:

    This might have been answered previously, but I remember that during November there was hope that A21 would get released before Christmas. That obviously didn't happen, but we're close to being in April now. 

     

    Without going into the "when will you release the kraken" discussion, could you guys explain what lead to that big of a planning delay ?

    Features not being as far along as we hoped. We don't primarily schedule features based on time, but functionality, so it takes as long as it does for that functionality to work. Many devs would cut features to hit a time slot. We take the time, within reason, to get the features we want.

  20. On 3/20/2023 at 4:51 PM, Thug Hunter said:

    @faatal

    Could you explain this to me? One thing I noticed about 7 Days is the fact that turning the graphics down usually doesn't have an impact on performance on high loads such as lots of zombies or uniquely place bocks placed by the player or removed. Matter of fact, my GPU would be more utilized by the game under less stress (less blocks and zombies) which would yield more frames despite the graphical setting I was at. So what I assume is the rendering loop is the same as the game loop for calculations? So the game won't render until its done calculating all the blocks and zombies? Cause it doesn't matter how much I turn down the graphics or how low resolution I make the game. After reaching 40 zombies or destroying and placing lots of blocks. I would get a performance decrease despite the graphics settings being high or low. Or it would be very minimal that it would not make a difference. Also the game seems to do fine with 30 zombies, but anymore zombies give exponentially less performance in frames, and some zombies would not respond to the player as if there is a hard limit. So is all the zombie pathing calculated separately or together? And if all these cases previously shown has something to do with how the game was initially made, how hard would it be to rework how pathing is calculated and how the game loop works. I don't know how your games works fully and I could be wrong, I don't know much about unity or use unity. So take this with salt because the point I'm trying to get is the less hardware your game needs to run, the more people can play it, but that also includes the ability for the game to utilize all of the processing power of the hardware.

    Many interactions between game code and Unity have to be done on the main thread, so that often is the FPS bottleneck regardless of what the GPU is doing.

    Zombie pathing is complex. There is shared data calculated around each player or group of players and then that shared data is used by each AI to find a path, but that step runs on one or more threads. A21 has changes to throttle the path request rates based on AI distance from players, but is buggy and not done.

  21. On 3/17/2023 at 5:32 AM, Crater Creator said:

    I don’t think it’ll hurt anything to confirm that the new forge has dimensions of 2x2x1. The devs are well aware of how this means layout changes to rooms with forges, since the level designers will be adjusting POIs that have forges to work with the new size.

    Yes, it is currently 2 2 1 and I don't know which way it extends.

  22. 5 hours ago, Rabbitslovecactus said:

    @faatalQuestion for the game when it goes gold. Will the sandbox mode have a setting to disable bandits if we'd like to? 

    We won't be deciding things like that until the feature is in and we actually play and balance it.

    2 hours ago, playlessNamer said:

    Will the ugly gap between sand/gravel and stone/concrete/wood blocks be fixed in the next patch?

    No. I said previously that terrain diamonds can't fill a cube exactly.

    2 hours ago, Trankitas said:

    Tomorrow is this thread's birthday, any info on A21 to celebrate? @faatal

    Sure, each day is a day closer.

    17 minutes ago, Chris Dunaway said:

    Does this mean that the road markings will not be a block anymore and will be painted with decals instead?

    No. Different systems. A hand placed block on a road is not a decal mesh generated above a terrain block.

  23. 9 hours ago, khzmusik said:

    @faatal Apologies, I looked back through the thread and didn't find it, but I believe you said you were working on some kind of "decal" system for things like bullet holes? Or am I imagining things?

     

    If you are - is there any chance this same technology could be used to make footprints in snow? (I've been watching playthroughs of Sons of the Forest and really like how that game does it.)

    I don't remember, but I was and am working on the decal shader, which is not a systems change. Just a change to rendering the decals we have on roads.

    1 hour ago, Blake_ said:

    Nor does (AO) render workstations nor trees properly. Sometimes they disappear when the area has a bunch of them. It has been reported ages ago but maybe what you did to plants could fix the issue. Or it might be related to failing raycasts that cannot reset|update themselves properly upon any kind of memory or CPU overstressing.

    AO draws shadows where inside edges are. It does not make things disappear. Those bugs are normally with the occlusion manager, which has some fixes in A21.

×
×
  • Create New...