Naz

Usually decaff coffee and water, or beer with very occasionally shots of what ever spirits i have lying around. Never been a fan of eating at the pc, especially sticky or greasy food. When i was a kid and you have pizza and are taking turns playing RE4, no one wants to be handed a controller covered in grease😄so that's always stuck with me. Although very very occasionally i'll eat something that can be eaten without touching the food, like sweets from the packet etc. Maybe i'm just a weirdo😂

Unfortunetly I haven't used a mac since like 2001, so I have no idea how you would do it on a mac. Looking at Google seems to suggest it is possible, but might require some c++ or python. However for all I know it could be done as easily as Windows, I just don't have any experience in mac os.

Yes settings matter when talking about performance. I usually test 2 sets of settings, everything cranked and one with lower settings in key areas. Reflections will tank performance the most , quality, reflective shadows and screen space reflections can severally impact your performance. shadows is next but not as bad, then options like tree quality is an odd one as it seems to control how close high quality block models start rendering, set it to low and you'll see trash piles popping in everywhere 😛 but it can help to turn it down to medium or high. But yes the discussion was about being GPU limited not fps limited, the distinction being if you're gpu limited it means there is a bottleneck somewhere in your system and you're not getting the gpus full performance. This doesn't mean it will be a bad experience, it could still give perfectly playable framerates with settings your happy with, if that's the case then you can just enjoy the game and not worry about it. However if you're not satisfied with your systems current performance, step 1 is to find which component is the weakest link and if you already have a powerful gpu it's usually the cpu.

I've tested the ryzen 3950x, i7 6700k, i5 4670k and am now testing the ryzen 5950x The 3090 tests were done on the 3950x, which is a touch slower than your cpu (In 7dtd) I haven't personally tested the i7 10850k or the 6700xt, so I can only estimate. However I think your cpu should be able to keep a 6700xt fed at 1440p, I wouldn't imagine it would run at lower than 80-90% gpu usage. But that's just my guess based on benchmarks on those components, when you throw it in there let us know how you got on👍

haha no worries i get that, sometimes you just have time to skim😛

Yes out of the box you would get better performance because your cpu is already running the game with 4 cores since your i5 has 4 cores. However newer hardware can get the same 4 core advantage by simply assigning 4 core affinity and since the newer hardware is more powerful, it will perform better apples to apples. case and point another 4 core cpu

Read the post above yours and the magic behind the curtain is revealed XD

yeah the 7dtd community are a great group, happy to welcome you aboard😜I hear that quite a lot, "my old hardware ran 7 days better" and they're not wrong, but once you apply the affinity tweak you're back on even ground and then the newer hardware is surprise surprise faster. I think it's because unity likes 4 cores and for a really long time intel 4 core i5's were the choice for gaming. Fast forward a couple years and 16 cores are available on consumer chips and unity just doesn't react well to having so many threads available to it with modern cpu's, especially with ryzen cpu's that have multiple dies, you want to keep all the process for 7dtd one 1 die for the lowest latency. Intel still benefits from the affinity tweak but not to the same degree. If you have the budget for it and are happy with it, then sure ryzen 5000 chips are the fastest gaming cpu's available today and 12 cores could be put to use in other games and certainly future games. I certainly understand wanting to get the most out of your system, especially with the prices of gpus that's been going on, gotta get every pennies worth out of it But keep in mind even ryzen 5000 probably won't max out your gpu at 1440p, but you will be getting the most fps possible at 1440p (excluding overclocking and faster ram)

There are some other tweaks you can try, but they're more involved and yield not that much more performance. I've never been able to break 60% avg usage on my 3090 at 1440p. So it's important to understand at 1440p in the current alpha 100% utilisation isn't possible. As long as the performance your getting is enjoyable to you, then you don't need to doing anything more. 1. Create a custom stripped windows 10 install iso This can help a little with your 0.1% lows, but on a modern cpu they'd yield 5% better at best. 2. CPU Overclocking This can maybe give you a couple % more gpu utilisation. 3. New hardware swap your 3200mhz ram for 3600mhz or drop in a 5700x, that will give you around 20-30% more performance (just upgraded to the 5950x so i've not run it through my usual suite, but that's from initial testing to give you an idea). But i totally get that the 3700x isn't that old and 3200mhz ram is perfectly fine in most games, but it is an option. Upgrading to a 4k monitor as meganoth mentioned would also see much greater usage of the gpu. I benchmarked many of these tweaks with my old 3950x, so it should give you an idea on what to expect and if it's worth it for you. Link

Agreed, it's not impossible to get gpu bound in 7dtd. However many players will run at 1440p and 1080p to get higher refresh rates over resolution and it's those common resolutions that are more likely to be cpu bound. I've tested the 3090(cpu bound at 4k) 1080ti sli (cpu bound at 4k) Single 1080ti (cpu bound at 1440p) 1080 (cpu bound at 1440p) 980ti (cpu bound at 1080p) R9 295x2 (been a while can't recall) 1060 (been a while can't recall) 750ti (been a while can't recall) So while some older cards could benifit from dlss, they're to old to support the feature. Modern cards that do would need to be running at 4k to have a chance at any performance benifit and sometimes even then it will still be cpu bound at full Res. I've seen a couple cases of the 3080 reported similary as the 3090 which makes sence since they're not far off each other in terms of performance. I haven't tested lower in the 3000 stack or anything in the 2000 series, so I can't say for certain. But it's my opinion the cases where dlss would benifit people would be in the minority. I think some people would turn it on and play with slightly worst image quality, when it would be unesseseary in cpu scenarios.

7dtd is a heavily cpu bound game, the extra performance of solutions like dlss come into play with the gpu not having to render the game at full Res. So it's benifit in 7dtd would be limited at best. Also dlss is only available as far as I know on 2000 and 3000 series gpus, so if you have a modern gpu you're not going to benifit from the performance uplift of having the gpu do less work, since mid to high range gpus don't even get full utilisation at full res as it is, due to being cpu bound.

Yes 30-40% means 2-3 cores at 100% 7dtd will usually max out 1 or 2 cores and that's where the cpu bottleneck comes from, even tho you still have 60-70% Total cpu utilisation unused. Happy to help😃

It is 100% CPU bound. Modern high end GPU's will always be CPU bottlenecked , regardless of CPU Model especially at 1440p. However you can increase cpu performance several ways. 1. Assign 7dtd 4 true cores This seems to be quirk of unity in general, It scales in performance up to 4 cores, then the more extra cores it has, you get an exponential loss of performance. You want to assign "True Cores" not virtual cores (Hyperthreaded cores or smt for amd) your real cores are always (at least in windows) going to start with core 0 then a virtual thread (core 1) then a real one (core 3) etc so forth. There are numerous ways to do this, 1 is task manager but it won't work running eac and you have to apply it every launch. The easier way of doing it is to set up a basic shortcut to launch the game with the desired affinity. I have 1 already set up if you want to just use that i'll link it here it's setup with the default install path, so if yours is installed somewhere else you'll need to edit it to point to the 7dtd exe. there are 2 shortcuts (1 for eac and 1 for without eac) or if you'd rather set it up yourself there is a guide here This tweak will give a quick and immediate boost to cpu performance, but i can't promise it will be enough to get 100% utilization at 1440p. 2.Make sure in your motherboards bios your rams xmp profile is enabled Ryzen love fast memory and if it's running at stock, your cpu's performance will suffer greatly. 3. Close or assign affinity to background programs if it doesn't need to be running, close it. If you want to have it running however. You can use the 1st method mentioned but assign the background programs the other 4 cores on your cpu. Doing this ensures your background programs will have much less of a impact on 7dtd performance when running on different cores. However it's still not as good as not having them running at all, so make sure your only running what you need. 4. adjust your video options Some options are very cpu heavy. Disable screen space reflections (checkbox on the bottom right) change reflection quality & shadows to low, disable reflective shadows. Set tree quality to medium or high. The other checkbox items will also help a little on the cpu godrays etc but don't make as big a difference as the other listed, so up to you on those.

Oh i did, neither of your linked "sources" are from the unity docs . They're troubleshooting form posts from unity devs and if you read the posts as you suggest, they don't support your fantasy either, everyone in those seems to think it's not normal and to contact unity support. No one else there seems to have the same problem, just as no one else with a high refresh rate monitor is having your experience. I can run 7DTD, Subnautica, Subnautica below zero, all with gync and vsync on with well over 60fps. What your experiencing isn't normal and getting mad at someone trying to help you correct it isn't helpful either. I already have done my homework , i've benchmarked this game for hundreds of hours across multiple alphas and unity version, with various settings on half a dozen different systems. The only time I've ever had the game locked to 60 with vsync in all that time has been a setting misconfiguration. Don't believe me if you wish, keep playing at 60, but i'm certainly not locking my game 60. This topic has gone off topic too much as it is, so this is the last i'll say, enjoy the game have a good day.

so no source😆 what you're describing isn't a unity problem, it's a setting problem. GPU's don't render partial frames (Unless it's rendering with SFR in an SLI Setup but never independently rendering partial frames) What actually happening is your monitor with vsync off is displaying multiple COMPLETE frames. This is called screen tearing and happens when your GPU is rendering more frames than your monitor can display, or fewer frames than your monitors refresh rate. The reason your locked to 60fps in games with vsync on ,is because you've not noticed that your monitor is running at 60hz. Yes 60hz+ monitors don't always run at the hz on the side of the box. This happens sometimes when you have a driver update, plug/replug the display cable, change primary monitor etc so forth. You can once again enjoy 60+ fps gaming with gsync and vsync enabled, by opening nvidia control panel > under display > Change resolution > Refresh rate drop down

That's interesting, I've never heard anything like that before. Can you link your source, wouldn't mind a read?

I always sell anything I'm not using, for pc hardware I have pictures like you said, but I also keep the benchmarks if I ever wanted to compare. Older hardware usually doesn't have much value, although some older hardware is worth more than it's original cost decades later. I always try to sell as soon as I no longer use it, to get the most resale value from it. But some things are just unsellable, I've had books and dvds listed for years for pennies, no one wants them XD

I've wanted to do a large test like this for a while to see how each tweak or overclocked component contributes to the final "Maximum Performance" results. Some interesting and surprising results all said and told. The way i've tested everything is as follows. All results have a "Max Performance" table always on top. Below that are the results from disabling or removing 1 individual tweak or overclocked component, to see what impact they have compared to running all tweaks. On the left is an overall difference table that adds up all the results from a data category like "average framerate" while doing the same for the other table and then comparing the difference. The 1 exception is the custom windows ISO install, this is just my usual benching methods without some of the rest of the tweaks, It's also done in A19.0 instead of 19.5 since i started these tests after i had already wiped the vanilla windows 10 install. I could have reinstalled vanilla windows 10 and retested and simply wiped again and installed the custom image. However i can't remember exactly everything i had installed and how every single thing was configured. So since i wouldn't be able to have them set up identically program and configuration wise anyway, i've opted to save 16 hours and do pretty much the same comparison but in a19.0. My Other 7DTD Benchmarks Benchmark Notes & Disclaimers Full Stock VS All Tweaks The first results are comparing a full stock config to running all the tested tweaks. As usual the original excel files are available Here. The other tested tweaks are as follows: - Custom Windows ISO (A19.0) - Assigning Affinity (Limiting the game to use only certain CPU Cores) (A19.5) - Having Background Programs Open (A19.5) - CPU Overclock Disabled (A19.5) - GPU Overclock Disabled (A19.5) - Setting the GPU to prefer max performance (A19.5) - Ram OC Disabled (A19.5) A19.5 Ultra Settings A19.5 Lower Settings Conclusions With Ultra setting we see an uplift of over +130% for the average and +100% for the 0.1% lows , Lower setting saw nearly 90% and 60%. This is quite literally double performance, no shelling out for new hardware, no game optimisations , no cheating lowering settings or resolution. Anyone can apply these tweaks, most of them are really simple and take very little time to apply yourself. However it's not all roses, some tweaks as you'll see do very little. However i've included them as they only apply if you have similar hardware. Depending on what exact hardware you have they might make no difference here, but could help substantially for your system. I'll go into more detail on these cases in each of the tweaks conclusions. For now this demonstrates spending some time tweaking and messing around is definitely well worth it. I'd go as far as say if your fortunate enough to land yourself some current gen high end hardware with the current shortages and scalping, you owe it to yourself to spend some time so you get the most out of your system. Benchmarks Assigning CPU Affinity Background Programs Open Benchmarks CPU Overclock Benchmarks GPU Overclock Benchmarks Setting The GPU To Prefer Max Performance Benchmarks Ram Overclock Benchmarks

Got a bit lax on tracking the performance of the last few updates. All caught up now, we have comparative benchmarks from A19.2 up to the current A19.5. To save time i just tested 1 system "Hotbox" since it isn't as powerful as my main rig it should in theory benefit the most from optimisations. My Other 7DTD Benchmarks Systems Tested Systems Overclock Notes Benchmark Notes & Disclaimers 2 Different video Settings Tested Hotbox Benchmarks Original Excel Files Download A19.2 VS A19.3 Benchmarks A19.3 VS A19.4 Benchmarks A19.4 VS A19.5 Benchmarks A19.0 VS A19.5 Benchmarks Conclusion The first few .xyz updates nothing much changed. However 19.5 gave us significant improvements and if we look at A19.0 to A19.5 we see huge gains in the 1% and 0.1%. The gains here up to nearly +30% is the equivalent to 2-3 generations of CPU single core performance, which considering this extra performance is not only free for everyone, it also requires the user to do nothing but let steam auto update. There wasn't as much an improvement in average frame rate but we still got a good up to 10% uplift. 10% doesn't sound like much but it can be the difference between unplayable and playable for some people. It's also important to note that the 1% and 0.1% numbers are arguably more important. These numbers represent how "smooth" the framerate is. Either a wide margin between those numbers and the average framerate or sub 30fps lows can show how often there is significant "stutter" or "micro pauses" that give a game an undesirable experience. Think if the framerate is consistently dropping to 5fps even if just for half a second, but your average is over 100 fps it would still feel unplayable. So in summery the pimps have done a great job here, if they can keep adding gains here and there in a20 and above, the game will run butter smooth for most people in no time. This is a promising glimpse into A20's performance also. It's not the end of the road for optimisations yet, but this is a good sign to come.🤘

It's a driver feature, the reason it's limited to certain titles currently is Nvidia only enables the feature on whitelisted titles, because it can cause worse performance in some games. So Nvidia only add games they've tested and found it benifits performance in them. However you can enable it for any game using profile inspector, the same tool used to enable features like sli in games that don't officially support it. https://www.google.com/amp/s/wccftech.com/heres-how-you-can-enable-resizable-bar-support-in-any-game-via-nvidia-inspector/amp/ I suspect the reason it doesn't seem to do anything in 7dtd is 7dtd uses very small textures. Most of the textures in the game are 2d and have file sizes of a couple hundred KB. Even high quality textures like the draw bridge are only about 13MB. So the current 256MB windows the cpu sees the frame buffer in is large enough already.

scroll to the bottom of the topic.

Resizable bar is a recent performance feature, available with current generation hardware first with AMD and then Nvidia and Intel. Up until now the CPU accessed the GPUs memory in 256MB chunks. Resizable bar allows the CPU to access the entire GPU's frame buffer at once, potentially improving performance. Titles that currently have this feature enabled can get a modest, up to* +12% uplift. You need 3 specific pieces of hardware to enable it. Also if you have a supported GPU it may need a firmware update as well as a motherboard bios update. Supported Hardware My Other 7DTD Benchmarks Currently the feature will only be enabled in select titles because enabling it in some games can make performance worse. However it's possible to enable it for any game using Nvidia Profile Inspector, which is what I've done here. Benchmark Notes & Disclaimers System Tested System Overclock Notes Tested With 2 Video Settings Trident Benchmarks I just did 1 profile for these tests to save time using G Profile V2. Custom profiles for games usually get reset to default after a driver update which indeed happened in the middle of the tests, but the RSB enabled profile was re-applied after the update and before further tests. So RSB should be enabled for all tests here. I also did the tests in A19.0 just because i already had the numbers with RSB disabled, cuz i'm lazy 😛 But i did do a 4k test in A19.5 for both settings just to be sure there hasn't been a recent addition that might benefit RSB. Original Benchmark Excel Files: Download A19.0 B180 Ultra Settings A19.0 B180 Lower Settings Conclusions So that was not great. Overall it's within run to run variance. However it's consistently an improvement, so that does suggest it does indeed make things better overall. Anything bellow 1% is almost certainly margin of error, so if it does actually provide a boost in 7dtd it's bellow 1%. I checked and re-checked to make sure that both RSB was enabled at a hardware level and the required bits in profile inspector were applied and indeed they were. So this means 1 of 2 things, 1. There is some other factor in play that does not enable Resizable bar in 7dtd or 2. Resizable bar just won't have any meaningful benefit in 7dtd. I even tried setting a custom resolution on my 4k monitor to run 7DTD at 8k to put more strain on the GPU's frame buffer. Even then consuming 16 and half gigs of vram there wasn't really any difference. If you want to try it yourself i'll leave the profile HERE but please run some before and after benchmarks to see if makes any difference on your machine, since the tests I've done show no real difference.

Are you talking about the ssao and reflection options? Or console command options like poi's?

What method are you using to assign affinity? You should see at least a 50% increase in performance.

Ryzen 3000 cpus won't ever be fast enough in normal situations to max out a 3090 in 7dtd. If you follow my previous discused "4 core affinity" advice, you can get a much more desirable 60-80% gpu usage at 1440p. Also 7dtd doesn't scale past cpu 4 cores, so a 12 core 24 thread cpu will never see 70%-90% usage in game, regardless of what gpu is installed. You must of had something running in the background like a steam game update or Windows update.