Jump to content

HeLLKnight

Members
  • Content Count

    53
  • Joined

  • Last visited

Everything posted by HeLLKnight

  1. Yes I get your point but I needed to use low with a high-end PC to get 1440p on high. Just did a quick test to 19.2 b3 and I can notice a very noticeable performance improvement with terrain quality on high compared with 19.1, huge kudos to TFP!
  2. Thank you very much Faatal, now that you're working optimizing terrain shader please take a look at this thread, there is a massive 40% fps drop from low to med terrain quality setting:
  3. A roadmap would be much, much appreciated! a lot of early access games are doing it like Subnautica Below Zero, Green Hell, Rust, Risk of Rain 2, Grounded, Chernobylite and many more... no need to be much detailed or exact dates of course. I love this game but don't have time to read hundred of messages here from Madmole, Faatal etc and I want to know whats going on. This also could good for sales like having steam page updated and keeps people engaged over time.
  4. Hey, just testing how much FPS eats terrain quality in a19 b180. You can find people talking about this here, reddit and steam. In past alphas, screen space reflections were eating FPS like hot cakes and devs addressed it. Now terrain quality is the major performance hog. My specs: I7 7700k, 16gb ram, ssd GTX 1080 8gb vram win10, latest drivers 1440p vsync off High preset except terrain quality Yes, there is visual improvement on clear mountains like desert, but do you think it justifies 40% drop in overall FPS just from low to medium? otherwis
  5. Hey @faatal and others did you notice this thread on Reddit? https://www.reddit.com/r/7daystodie/..._to_die_unity/ Looks important, some users are getting a HUGE performance boost easily. Maybe these Unity options can be set as default for A19 if devs find are stable enough
  6. Thanks MM, I think a simple roadmap would help with our anxiety a lot and would be very appreciated something like star citizen roadmap but much simpler: plain text, updated every 2 o 3 weeks like this: Graphics - Upgrade to Unity 2019.3 - Done - Upgrade to lighting system - WIP, 17 of 22 tasks completed Weapons and tools - New AutoShotgun, Machinegun, Steel Club - Done - New Junk drone - polishing - Retexture and remodel old weapons - WIP, 8 of 10 tasks completed Vehicles - Vehicle mods - done - New collision system - scheduled
  7. Nice to have AA options! thanks faatal. TAA mode? IMO the best, produces blur in some games but I love to combine TAA with that new contrast-adaptive sharpening options appeared in lasts months in both amd and nvidia gpu settings
  8. @Faatal We heard you since april 2019 talking about updating to Unity 2019.3, which means we can expect from 2019.1 stable Vulkan, performance improvements, better ambient occlusion, faster physics, final garbage collector and better tools for you, etc. So now that 2019.3 is final and looks like a19exp is not coming very soon, can we asume new Unity is coming to a19?
  9. Sorry I didn't understand, how could be this an improvement? Isn't supposed to be a new great feature in a17? Edit: in a17 exp b238 changelog says: Disabled texture streaming of atlases for CPU performance gains. Should I get better performance disabling texture streaming completely if I have 16gb of ram and 8gb vram? On other side: Can somebody better explain this?
  10. 8 hours ago, TFP updated /testing_trunk/ for the first time in the last two weeks with a big jump in buildID number So looks like they copied a build with big changes from /testing_branch/ to the more stable /testing_trunk/. Possibly they need a few more builds and internal testing to go to public experimental but hey, looks like we are not very far!
  11. Let's read: They talk about render path, so for me this means ingame, not generation. Also the performance improvement comes from a dramatic reduction in "Draw calls" which is one of the most expensive cpu tasks at any game. If you have to many draw calls, you end very cpu bottlenecked. And we know 7d2d is cpu bottlenecked for sure, but only a dev can tell us how important would be to reduce terrain draw calls in this particular voxel game. Even if 7d2d is not pushing draw calls too much now, as far as I understand jumping to Unity 2018.3 would allow devs to put a ton more objects on s
  12. I really hope they are trying to update Unity to 2018.3 because new terrain system sounds too good: https://blogs.unity3d.com/es/2018/10/10/2018-3-terrain-update-getting-started/ They claim more than a 50% reduction in CPU costs plus big improvements in quality before-after, identical triangle count:
  13. They have been very busy in the last day according to steamdb:https://steamdb.info/app/251570/history/ They usually use /testing_trunk/ folder to test before updating /latest_experimental/ folder (public experimental builds) but this time you can see a ton of updates to /testing_branch_alt/ which could mean they are testing some big system change in a separated build. My bets are new RWG algorithm which can break a lot of thing vs the old system, or new unity version (they talked about adopt new garbage collector to fix microstutters), new occlusion culling system which avoid render thi
  14. Alpha 17.1 exp b9 patch notes: Looks like TFP is very happy with 17.1 exp balance feedback from us and they are working on major things, hope they fix RWG and make AI less exploitable in blood moons soon. Time to get an updated 7d2d roadmap for short/med/long term? no dates ofc
  15. Faatal, just seen 17.1 horde night from Grumbul's twitch stream, unfortunately this kind of exploit continues to work making it boooring: I've read this on 17.1 patch notes: AI will do destroy area 60% on a fall after unreachable side path. I expected some zeds cutting the loop and start destroying pillars but didn't happen in all the night. Maybe the zeds don't consider this unreachable even when they jump forward in the last section. Awesome job on this new experimental!
  16. Awesome new shots Dust2death, can't come soon enough! Maybe it's me, I find some tree trunks a bit too bright compared with the leaves, or maybe it's the just the sun position? On the other hand, I think it would be worth trying to improve the appearance of the 2D grass on the ground, since you're watching it all the time. When the ground has the same color and brightness/contrast as the grass it looks great, but otherwise it looks pretty fake. Hell, I would even sacrifice a few fps for better grass. Rust game uses Unity and looks much better, do you think could be possible in this voxel
  17. Awesome mod Dust2death, the only downside I can say is that the modeling of the spruce trees is too dense and it is difficult to see through especially when driving. Maybe if the lower branches started higher in the trunk it would be better. Keep up the great work!
  18. Incredible mod! not only has better graphics but also gives better performance for me than vanilla. Also Unity 2018.3 gives "GPU-instanced render path for terrain, yielding a dramatic reduction in draw calls for most cases" as I understand it's important for open world games and moves calculations from CPU to GPU, crucial for this game.
  19. Different zombie behaviors can't come soon enough Think about a night horde like this (just few mad ideas): - Some different ferals with actual a17 AI or even smarter, but they doesn't destroy blocks at all - Others more stupid that doesn't know how to jump obstacles at all, and keep destroying pillars but they do less damage to blocks - Some even more stupid that are like a16 zombies taking out running in circles, not paying attention to weak blocks and doing more damage - Feral dogs and wolves that know how to jump 1 or 2 blocks FORDWARD when they see a hole - Do you like to pass
  20. As I guess this was discused many times and I don't want to mess this thread more than necessary, can anyone point me to some dev posts explaining why update cycles have been so long compared to other early access games? thanks in advance
  21. A17 changes directx9 to 11 and new unity version, let's hope it will help a ton with that problem.
  22. Can we expect that first experimental version (even only accesible to selected youtubers) to arrive before the HP bar empties?
×
×
  • Create New...