WIP Beta released A new *CITY* map - LOS INJURUS 2023-12-20

So is this supposed to be fictional? If you want I can try to think of a few names?
I feel like the town should be Bringon or Brindon or Bringeon? (I don't really know just think it sounds good)
Also that highway system looks great.

 

They sound like British fake town names TBH.

 

i see, the classic authors response on every beamng related mod.

 

I can't make any promises to release (besides that not being my words 'when it's ready', though I'm by no means upset, it's about as close as I could say anyways). Due to the unpredictable nature of development, I can never promise these things. It's look like a fall release date currently, for anything with a lot of meat to it.
So entirely, not sure what's in-store for release date. It depends also on how well people back the project to show appreciation for the work too (as with funding I can buy models and not have to make everything from scratch 100%). So I can't make any promises. It'll still come out regardless. However, one person making a whole entire city without using assets belonging to others (aside of BeamNG assets), takes a LONG time (which is also why I re-used one of my old end-of-life maps which had this plan for it anyways).

So this is a good time to mention Patreon supporters will get it early, this is especially beneficial for early reviews or crash compilation use. It's a lot of work on my end, so it's somewhat understandable that it not only takes a long time and energy (and costs me money!), but that I'd just like to cover my losses on it. I won't stop mapping if I don't, but still, It'd be nice if I could cover some of the losses.

What's this project cost me thus far?
Maya LT monthly subscription for ~30$ a month (3$ tax) for several months already...
Ram upgrade - 110$ for used 16gb kit so I could double up the RAM in here as I hit the limit of the 16gb I had in here DAILY
Had to buy an additional 500GB SSD for project asset storage, so I can organize better and not be constantly moving (and losing) things on the computer. Got this on sale for 87.99$ + 10% tax for a 500gb MX500. I don't care if I recoup this cost, as it would have been necessity anyways sooner or later.
Upgrade OS To Windows 7 Pro from 7 Home Premium, as 7 Home has a 16gb RAM limit (what I had previously). I don't care if I recoup the 50$ I spent on the Anytime Upgrade for Windows 7 Pro, That was my option, I could have done the Windows 10 upgrade, but after having to fix everyone's Windows 10 constantly, and the fact that it wants your data, and always has GB's of updates seemingly weekly (over my super-slow net connection!), I absolutely will not deal with that here one bit. Maybe on the next computer but certainly not this one. Windows 7 works great, so that's what's staying on here. If it's not broke don't fix it. I didn't have to lose any *ANY* of my settings either, that was worth the price of entry alone. I must say the Anytime Upgrade was painless though, and only took 5~10 minutes. Took me longer to install the 2nd pair of 8gb sticks into this machine (due to giant CPU cooler - because I didn't have to remove it) than it did to update the OS.

So, everything included, 55$ for new (tax incl) OS upgrade, 110$ for used matching RAM, 95.76 for SSD (tax incl, est), comes out to around 260$, plus about 120$ as of today for Maya (plus 12$ tax), Ouch 292$. That was at the point I decided to open up a Patreon account and accept donations, as this cost me a pretty penny and I am on a fixed income (not broke, just not rich, I've invested most of my surplus monthly money, for my future, also why my tower's four years old coming this October and I only chose to update the memory/OS/add an SSD vs build a new one).
Yes, this project has cost me just shy of 300$ USD. So if anyone has a few bucks to spare, see the 1st post and head to Patreon. Those donating north of 10$ USD will, in the short-term (a few days to a week or less), receive a link of my most recent map backup, which is done weekly/bi-weekly generally depending on workflow. I don't have charge-on-the-spot enabled, so money comes out at the 1st of the month AFTER you donated, ONCE, it's not recurring.

Regardless, I love to mod, and this will be a city full of parodies of business and town names, and as such, should keep your both entertained and chuckling occasionally too. So I'll be modding regardless. So those of you with no money to spare, or on a fixed income yourself, don't feel too bad if you don't have money to spare, or even if you're too young to have ability to spend online. There's no REQUIREMENT, There's no hard feelings what-so-ever. Early releases will still come out for free public consumption either way, so no worries.

That is all!

 

This is looking Great! Where is it based? I see palm trees so somewhere tropical? But sand?

 

Howdy folks, come on down to Chicken Munch, where it's B I R D M U R D E R T O - G O

 

Bottom left store

 

Well I have a name in mind, it's not these names above, but I'll throw a hint.
The 'county' it's in is going to be a play on words of NOGALES county (which I think is Arizona or Nevada or something, pretty sure it's Arizona). This should come as absolutely positively no surprise given the tongue-in-cheek humor in my other maps. Trying to keep it pretty G-rated though.
So yes, entirely, this map is going to be good - AND funny. Should have some new mapping pictures up sometime over the next few days, after taking 3 days off for good measure from mapping/modeling stuff so I don't burn out ...
(and also, to do my lawn, repair the deck [because I don't fancy going through it], clean up the house and such, that's all done BUT I still have more stuff to do like paint some door frames, get up on the roof and check that out - there's about 4100 sq/ft of it too ARGH, do window/door caulking and such)... Real life sucks sometimes like that but if I don't do it, this place will fall to pieces.
What to expect?
Large commercial buildings, medium and high-rises.
Parking decks - because YES we NEED those - and of course the tall ones will have a jump or two near the corners so you can fly off, because driving all the way back down is booooooring. It's like 'well why did I go up there in the 1st place?' ...
Possibly a new subway station building, though those might come later not sure.
More houses, because I'll need another 60 or so more before I start fleshing things out enough to get some variety, we don't want every neighborhood looking 'same-y'.

--That's all, keep the questions coming if you have them, suggestions too, and don't forget the Patreon page link in the 1st post if you wish to help out the project! Thanks to those folks who have already donated to the project.

Also, Nevada Interstate is *NOT* broken. It works fine, must have been a fluke or was patched-fixed by the developers.

 

Will it be FPS friendly like the Roane County TN map?

 

I am working on FPS to keep it better. It's too early to tell, but it should be around the way West Coast is, in the most developed areas. It will be easier FPS around the large commercial buildings then it will around the houses which have plants/trees which eat FPS for breakfast.
Trying my best here. I have a 4km visibility in this map (it may come down to 2), so it doesn't usually drop below 30~35fps in finished scenes with medium reflections on, on an RX 480 with the details otherwise MAXED, so that's pretty good. Turning reflections OFF keeps things above 40fps nicely, usually 55~60fps or much better. Rural areas with less homes will have much, much higher FPS.

I am doing whatever I can to make this best for everyone.

 

West Coast USA, unfortunately gives me like 5 or so FPS throughout most of the map, but I can always lower the graphics. Roane County gives me around 20 FPS (with Normal graphics). Map looks awesome so far by the way

 

Okay, Well, it should perform better than West Coast USA, because West Coast is all one BIG model for the driving spots (roads except rural ones, sidewalks, retaining walls, aqueducts, bridges, tunnels, highway), and the buildings in the city are one big model - all of them except some houses in a development on the far side of the highway, and the oil derricks that animate (and the race track stuff is separate). So yes, entirely, that may be what is overwhelming your gpu.
This map is not all one model - it's the exact opposite. It's a modular city kit - which will be free (early to Patrons), where the roads are all puzzle-pieces - like SIMCITY 4 NAM!, if you've used it. This way only what is within ~320 feet or so (100 meters) is displayed in it's full detail. Beyond that out to about 500~600 feet (200 meters) is 'half detail' meaning half the poly surfaces or triangles (these are the building blocks of all polygonal objects in 3d games beyond 'half life 1 and 2', 'quake' and 'dos doom' - which were BSP based). Beyond that at around 1000 feet or 320 meters, you end up seeing something close to the LOWEST 3d detail levels, which are 3~12 poly surfaces and/or triangles, depending on the object, beyond the most objects go 2d billboard sprites, and when you're a half mile or so away, many buildings and roads won't even render at all. So in-stead of displaying on 100,000 triangle or poly-surface model (ugh) which is fine by today's current generation cards, but NOT for something older like a Nvidia 550ti or 650ti or Radeon 6570 1gb, you only display the surfaces needed, and this (unlike the 'all as one model' method) can be adjusted with the MESH DETAIL settings much more natively.

I would like your system configuration, as complete as you can list it, especially, processor, graphics card (and VRAM amount), system Ram amount (and speed if you have it), and if you have the OS and game installed to SSD or are still using a hard drive. This helps me better understand who I am making a map for.

Thank-you for your interest in my map project. This is my most serious project yet in 24 years of game modding.

 

I can't really remember any map which would of been problem to run in terms of GPU power, even with 1050Ti reducing graphics details a bit, hardly noticeable by my poor eyes, did the trick.

Problem always is single core performance of CPU and you can't do much about it (except get that 8086K and overclock it), sure you can turn shadows off which helps a bit, allows more room for collision surfaces, UI, etc. on that same CPU thread, but that causes huge change in visual quality.

With Roane County, there is practically no difference in low FPS between gtx1050Ti or gtx1080, also as I run 60fps maximum FPS, there is not much difference in average FPS as only reason for dropping below 60fps with 1050Ti is CPU.

Hence I think that optimization of maps should be focused on optimizing CPU usage, while GPU usage being also important as can be seen how Roane county delivers most of the time 60fps with 1050Ti.

West Coast USA is quite GPU heavy, but CPU is still major limiting factor on that map.

6700K would be much better than my something like average gaming computer level 6700 in this regards, 8080K would be top of the line.

I keep hearing about UI overhaul and that I look forwards a lot as UI is one thing that takes a lot of CPU cycles currently and on same thread as collision surfaces and some of the graphics are running, hiding UI gives a lot of FPS boost in situations where I'm dropping currently under 60fps. (I will usually not play areas of map that drop below 60fps, it just makes driving a mess, FFB for example will not work properly below 60fps so my goal is always keep 60fps and not to have any drops below that).

Yesterday I figured out that prefabs (buildings) in Glass view mountain map (autumn colors version) has most effect on FPS, while mesh road objects being high impact, those buildings were one that brought my FPS to 30 on that map. Again CPU is the reason, too much workload.

Question is then, is many objects with collision surfaces worse than one object with less collision surfaces?

I think that without WCUSA being one object, CPU load would be higher, because while you can LOD graphics, can it be done to collision surfaces?

Factors that make CPU load too high are complex combination of this everything and sometimes optimizing for GPU can lead to de-optimizing for CPU, something that should be kept close eye on for sure!

GPU load of 30%


So you say dynamic reflections, ok let's turn them off, however we still get 30% of GPU load, CPU gets bit easier time so FPS is bit higher, but it is still CPU limited situation (don't believe CPU load percentages being absolute, it is more of rough guideline that one needs to interpret), it is about same if it is 1050Ti or 1080, even iGPU gets near the same, although it can't do that on high graphics of course:


Thus, while GPU optimization helps on some areas, those collision surfaces take up available CPU headroom, which so many other things need, CPU render is very high on built areas, no matter what map it is, so I'm afraid that focusing too much on GPU optimization might have less FPS on slow spots than choices that make map less GPU friendly.
Also it might be that it leaves less available options for user to effect on their FPS as affecting CPU load has so limited options available, while affecting GPU load there are more effective options on user's side.


So question that should help some is, when you get 30ish FPS, how far from 100% is your GPU utilization in that situation?

 

I am *FULLY* aware of CPU load in my Roane County TN USA map, it's much too high in built-up areas, but everything there is separate. Even houses may be multi-piece. The trees also eat lots of load in both GPU / CPU (and also RAM footprint - collision eats up lots of RAM footprint).
However, this map is different, it will be balanced as best as possible (by my RX 480 8g GPU and 4790k 4.4ghz all-core speed CPU w/2400mhz CAS-11 RAM - this helps by 10~20% over default Cas 9 1600mhz RAM FWIW). It's still going to be a bear for machines meeting bare-minimum requirements without resorting to most basic settings, but that's not my 'fault'. If people want things to run on Potatoes, play GTA SA, that game is fun and might keep you occupied longer than my map, surely. However, for the here-and-now and the last few years of PC tech (my machine's almost 4 years old), there's my maps, to stress-test your machine, no matter HOW good it is, even a 2000$ 7980XE will close-to bottleneck on single-core performance depending on how you've tweaked it and what RAM you have, and what GPU you have (in this example, the person had a 1080ti).

So yes, yes, It's actually better off and much faster to build the way I am building (read: lots more stuff built in the editor not in the modeling software which I cannot test within it), but there's trade-offs. Too many decorations and the map will chug on the game-engine feeding the GPU/CPU/RAM with the appropriate things to render.
There was a bit of performance regression with 0.13 on my maps, as much as 15~25% speed in spots, with the thread scheduling, but this will disappear when running lots of vehicles, especially with LOTS of cores on the CPU (which I haven't really tried much yet to be honest).

The 6700 i7 cpu won't even run the game faster than this i7 4790k, and may in-fact be up to 15% slower, due to lower clocks on a newer process. The 6xxx series of Skylake was really lack-luster for improvements, read: there weren't any short of DDR4 and DMI 3.0, but you're still better on power consumption then an older 4xxx series cpu would have been. Feeding it super-fast RAM will help some, but isn't worth the upgrade cost if you already have 16gb of RAM.

Building a RIG for BeamNG ? Buy a K-series CPU, especially if you want to push the limit. 2 years after it's release people are still bench-marking new machines on my Roane County TN USA map - because it's over-the-top. Not because I wanted that, but because it's the only way I knew how to build what I wanted with what I could.

More map objects > more use of CPU to schedule what needs to be in RAM, L2/L3 CACHE, Sent to GPU etc, to be rendered and used. LOD use here is very efficient, but one must be careful not to put too many things in scene. Moderate to heavy GPU use on very full scenes, very minimal GPU use otherwise.
Larger map objects made of combined things> Less CPU use but MUCH more GPU use, may encounter micro-stutter or just plain LAG when caching very large objects both at map load & during play. These super-large objects will fill in lots of the screen and be seen almost all the time causing LOD's to not get used much.
Map is one big object > Don't do this. Map will run like butt because the object will never be out of view and will display 100% at all times. Both CPU due to collisions on such large object and GPU due to rendered poly's will be maxed out.

So I tend to fall mostly on the first object. It's most responsive to CPU and GPU increases in performance and scaling. I did a lot of tests, and it also offers me the most freedom - in something I didn't cover.
If you make stuff as all one object, your VRAM (visible model), System RAM (collisions) have to all STORE, PROCESS (via CPU) and then RENDER (on the GPU) all this stuff SO many times. How many times are we going to render that same bush full detail, that same section of road or sidewalk, that same pillar, etc - and store the redundant collisions for them all.

So if you can line everything up nice (takes a little time, made up by the ease of placing and duplicating things in the editor when needed), I find that the puzzle-piece approach is best, even if it does tend to hog up slower cpus.

For what it's worth, the scenes with houses, city parks nearby too, will in-general be the MOST COMPLEX parts of the map, which is why I built them first. They have the most objects, the most vegetation, and keep the CPU and GPU busy. I built them first as a bench-mark of what to expect. Houses are 30~100 poly-surfaces depending on how complex they are (most are around 50~68) at full detail, where-as trees are generally like 500~600+ surfaces each - at-least! So I am definitely considering adding some more optimized vegetation into this map before it gets too far.
I have been trying to build some visibility blockers (OCCLUDERS) into hills but it's buggy at best with this game engine and if there's a road over it, your car shadow disappears randomly even if the blocker if 2~4 meters under you or the object you drive on.

TLDR= It won't be as bad as Roane County on the CPU hogging front if I can help it.

That is all!

 

Actually Kaby Lake was even worse, it used higher clocks than Skylake, used more juice and had not much overclocking potential without delidding.

Since that 4790K there has not been much of any improvement though, all CPU's are really poor options, while Coffee Lake does improve performance a bit, it still is just a mere 11% more than what you have:
http://cpu.userbenchmark.com/Compare/Intel-Core-i7-8700K-vs-Intel-Core-i7-4790K/3937vs2384
CPU is over 400 alone, add MB etc and no point in upgrading, even from your not so old rig.

6700K is about same with your CPU, which really is pathetic as is single core performance of 8700K, so many CPU generations and still no good improvement to SC.
http://cpu.userbenchmark.com/Compare/Intel-Core-i7-6700K-vs-Intel-Core-i7-4790K/3502vs2384

Ridiculous was also decision to downgrade iGPU from what they had before Skylake, but then again we have been really long at point where there is no point of upgrading and it does fit the theme. So before 10th generation, I doubt there is any reason for either of us to upgrade for any good performance gains in single core and by then there is probably again new memory type needed, so in this era machine that is 10 years old is still quite recent in terms of performance, such poor has been improvements in chips.


Now trouble is that your CPU is something like 190 or better in Cinebench score (I believe 8086K is 219 or so), but majority of users don't have that kind of computing power, majority of users have probably that 20-30% less single core performance, which can make it more challenging to estimate what kind of performance most of the users are getting.
You can see on forums how people complain about WCUSA performance, for me it is about just and just keeping 60fps with best looks, so it is good to know that you aim for better than WCUSA performance, but if you aim CPU limits of your CPU, that is going to be only few users that can run the map without dropping frames.

From Steam users that would be those who belong to ~7%, average BeamNG user might have bit stronger CPUs, but still majority is having lower SC performance than you have or even what I have. Yep, people run this game on potatoes and some run it even fine


However as UI updates is taking quite large part of the that CPU single core computing power right now, it might be that rumored UI overhaul will help somewhat on that department.

For me 6700K would be helping quite a bit of single core performance, but then again there really is no other games than old Arma Gold along with BeamNG that actually would benefit from such upgrade and it is quite questionable that situation will be such forever with the BeamNG. With Arma Gold, well, poor optimization of whole Arma series is a running joke among gamers.

In latest builds of BeamNG, it seems hiding UI does improve FPS a lot, so something there is definitely taking up bit more of CPU time now than before.

I am really curious what master builder's map will be when it is complete, I'm sure that it will be awesome fun, so I am looking forward of it!

 

Good to know there isn't much out there for upgrading, that's why I boosted the memory and switched from 7 Home OS to 7 Pro OS, and threw in yet-another-SSD. I wasn't going to put out a ton of cache for a machine better than 98% of folks, to help me make a map only I could run well. Even if it was 200$ to re-do the whole machine I'd likely still not do it. When AMD has a 6-core APU chip, that has awesome integrated GFX performance, and there's DDR5 system RAM, I might consider upgrading to save some on the power bill.

I have been working a lot on the map objects lately. Getting all the LOD's done on the surface roads, this helps performance HEAPS, a lot more than expected, 20% in some places. Got most all of them done, just fixing up a few of the dark spots on the edges of the small filler pieces now and that's it. Just a few pieces where driveway aprons are (the ramp on the sidewalk so you don't have to drive-over full curbs to get into a driveway), and all will be done with the smaller surface roads. There's lots of these so that should buy another 2~3 FPS in very developed areas on top of the 5~7fps I gained back yesterday. No visual degradation what-so-ever from better LOD's either.
Another boost to performance is improving the quality of, and fixing the bugs in, the Rail pieces (ground level ones, the subway ones have LOD's and are 100% working). This will also help performance substantially where they're used, maybe 5~8% (that's substantial when it comes from one set of objects). Since someone was nice enough to send me a working train, I could properly test them for the 1st time, and found a few bugs within a few pieces. Thanks for that!
Long story short, I am working more on optimization now, so I can get this city kit ready to ship for Patrons near the end of the month. The rail kit by itself will hit the forums free for everyone to integrate into their maps, once I get all the bugs out out of it and have it optimized nicely.
Lots of work to be done yet on the curve pieces for four-lane roads though (the 5 and 6 lane pieces already have everything done).
Once this is done I can move forward mapping, and modeling a queue of houses I've got backlogged on again!

--Cheers!


EDIT:
Long curving and straight track sections are being SCRAPPED from the project.

I've wasted hours messing with this, not going to waste any more time with it.
There's still plenty of good choices via the shorter curve sections and I *WILL NOT* re-make the tracks a second time. Just not going to happen, period, complete waste of my time when the shorter versions already have complete functionality in and of themselves. The rest of the issues have been fixed with the other pieces, though. 99% sure anyways. In due time I will re-make a long straight section but that's it.

EDIT again:
@synsol and @meywue The bugs struck again, pretty sure this is a beamng bug here and not my object.
Please have a look @ the attached model as to why a model (COLMESH is offset in the 3d modeling program, but matches in beamng), has to be modified this way to even work. Placing the collision mesh in the same spot as the visible mesh results in the above screenshot - which is bad.
Nothing is off-set from 0,0,0 that I can see in my model (until I had to move the collision mesh so it would line-up in game!).

 

Are there gonna be custom billboards as well?

 

Yes there will be plenty of those. A good variety likely just advertising my other maps, and fictitious products as in the developer's maps. There won't be any scantily clad women on them though, being that I am not a sex-starved teenager in mummy's basement

Seriously considering making one that says "Now Hiring: Scantily clad women to pose for billboards in driving simulation game that has nothing to do with scantily clad women! - contact Still_probably_a_virgin_@_mummy's_basement.yuk"

You know what, I'm totally doing that, bwahahahaha

 

I'm doing a little Photoshop project with a '52 Ford Courier (gonna make it a drag car). I could make that into a billboard ad after I finish the project.

 

Yes yes, feel free, I can throw that on a billboard no issue.

I will gladly host other modder's work, provided it's good work and contains stuff they've worked hard on making, vs some people just trying to port in GTA cars or something (not that I mind that, but if you've slaved 500 hours or just LOTS of hours on something, yes that totally deserves some recognition).

So yes entirely, 100%, I can do that. I don't mind giving back to the community. It will help with immersion too.
Not sure it will beat my Dishwasher advertisements "now featuring Spin Cycle, Sand-Blast, and the much coveted, TUMBLE DRY!"...
Never hand-scrub your dishes again