Why the hell anyone should care even he uses vob mods? How does it bother you or how is it any of your business as long as he don't share the stuff in here forum?
He has paid for the game and is allowed to do what ever he wants with it. Deal with it kid and stop crying about every little thing. I feel sorry for your mother.
Skirokko, You are not properly understood. Read more...
All vob mods have a low level of quality, slow down the work of the simulator, can cause problems in the work. As they are absolutely incompatible with each other, and some may even harm your computer.
AFAIK, many VOB mods change PHYSICS, which is bad for online racing.
The VOB mesh is not only a graphical model, it also controls collision detection in some way. Let's imagine two players using a VOB which removes all the collision on the car body. The two car may come so close that a car's front wheels are about to touch the rear ones of the other car. If the original VOB is used, the cars would be sky rocketing. But with the "mod", they can run with out problem. Not fair, right ?
Thera are 2 types of VOB mods. Ones does alter the colision mesh, so they are incompatible with online gamplay (now detected by dedi server since 0.6B). The others does not, so nobody will know you are using it.
Plus, VOB modes are in websites where cracks/keygens (so probably malware aswell) are shared.
Plus the fact that VOB mods are plainly forbidden by LFS rules.
As far as I can see, there's nothing special in that video. I'm pretty sure our devs already understand what's shown in that one. The problem for them now is how to code the physical progress into the virtual world without overloading our CPUs.
I know that Scawen is trying to reproduce as best as possible in the real time model the reactions of the complex not-real-time tyre simulator that he developed and is a lot more advanced than our current tyre physics engine.
It's all about CPU work-load, it seems.
If that's the case, I don't understand* why LFS code still use just 1 core.
Nowadays 2 cores are pretty much the norm.
With many people with 4 core cpus and even some with 6 cores!
* I understand it would take quite some time to "update" our current lfs code to use multiple cores but it will have to be done at some point, right?
Why not now? Since we're already stalling there's many years and I guess it would help immensely the new real time engine physics code to be viable...
While a change in the physical model may multiply / divide the CPU load by 10, 100 or 1000, multiplying available resources by 2 or 4 isn't of high priority.
Just my not-educated guess, could be simply wrong.
That is because the car in the back has zero air resistance thanks to the car in the front. Something similar happens in real life but maybe not so exaggerated.
Because the raw CPU power is not the main problem.
By using 2 cores instead of one you get only 50% of speed up, so if your algorithm is too complex, it will take instead of 1 month to calculate single frame only 15 days.
One problem is to design algorithm which is simple enough (in terms of computational complexity) to be solved several times per second (i.e. polynomial complexity O(N^k) at most, where k is nothing really big, like 2 or 3), yet the final results of such formulas have to resemble the true physics as much as possible.
Second problem is to get even those slow complex algorithms designed to run offline (ie. calculating single frame for minutes or hours) right, so their results model the tyre physics close enough to reality. Mind you, the real world tyre physics is not something you can read from the book and suddenly you have perfect set of formulas which do simulate it properly (if you have enough computing power to calculate them). Even tyre companies are still developing and improving different ways how to calculate "real world" physics, so they can even more accurately simulate their new tyre designs.
Once you solve to some reasonable level those two problems (Scawen is recently working on that offline full simulator, so the transition to simpler real-time version is yet ahead of him), then you can spend some time optimizing the resulting code to run not 20ms, but 14,5ms, and use maybe some additional core of CPU, but it's no big deal if the base is good. If the base algorithms are too complex, nothing will help you, certainly not two cores. So the raw CPU power is problem in last stage, when you are like 10% slower than you would like to be, but the main effort happens way before that stage.
If it was just down to zero wind resistance he'd still gradually decrease in speed due to friction on the road. But he's in neutral and his speed is matching the car in front, so the cars have definitely become connected in some way.
