Jump to content
  • Advertisement
DaTueOwner

Anyone who wants to write a little game engine?

Recommended Posts

1 hour ago, _Silence_ said:

What make these engines be different is not the fruit from hazard.

Of course it have different implementation of both components anf architecture . But it have a same set of components. If you want to have into you game, for eхample, chatacters you anycase need components for bones system and inverse kinematics into engine, independently is it only chatacter into your world, or it model any person into galaxy. 

 

1 hour ago, _Silence_ said:

Big games with very large world will use a lot of things to optimize, otherwise the game will be slow

It really has been actual 25-30 years ago for low-end solutions.  For example for tunnel walker/shooter you need to render a couple rooms that player can see at time, independently from total size of the maze. Othervice it really will be slow. With  open  spaces scenes it much harder to implement, but concept is same - it draw only little part of a big world at time. It why demos ever exists - it clearly shows what engine able to perform for world of any size.

Edited by Fulcrum.013

Share this post


Link to post
Share on other sites
Advertisement
2 hours ago, _Silence_ said:

Some of them have PhD, and this is sometimes required to ensure he/she would understand and be able to devise new

Of course, becouse phisics staffs involve same simulations that scientific software, with only difference that it optimized by speed (that much much harder to perform) instead by precission. Also if you ask any of its PhD what same stuffs involved , for exampe, to realistic looking water effect, he/she will answer that it is 99% of phisics stuffs and 1% of graphics stuffs. Really, developers (especially such high qualified developers as PhD) make line betwin stuffs by sciences that it stuffs involve, instead of gamers (and evet managers and promouters) that call a graphics stuff anything tha visible on screen. Really if you see a bar that glowe as metal "graphoniy", you never will say that it make from "phisoniy" metall covered by thin layer of "graphoniy", and end ever will have no any suspections that it thin layer consists of 2 sublayers maked from different isotops of "graphoniy". Really one of most importent part of graphics stuffs player ever never seen on screen, becouse it perform a different kinds of clipping algo, that determine what same required to render, and without wich gaming will be really to slow.

Share this post


Link to post
Share on other sites
45 minutes ago, Fulcrum.013 said:

Also if you ask any of its PhD what same stuffs involved , for exampe, to realistic looking water effect, he/she will answer that it is 99% of phisics stuffs and 1% of graphics stuffs, 

Yes because realistic light interaction requires simulation of physic laws and therefore graphics is involves physical calculation i.e. solving the redering equation. So I would say you can't really separate those concepts -> realistic looking water needs "100%" physics

Share this post


Link to post
Share on other sites
Just now, DaTueOwner said:

So I would say you can't really separate those concepts -> realistic looking water needs "100%" physics

Of cource rendering involve optical phisics, but it usualy part of graphics stuffs (shading algos) becouse its is same (ot near same) for any semi-opaque objects. Under phisics stuffs here meaned simulation of waves droplets and so on, that make difference from just semi-opaque plane that has simulate water for 20+ year old games. 

Share this post


Link to post
Share on other sites
1 hour ago, Fulcrum.013 said:

becouse phisics precission exampe betwin evet promouters tha glowe graphoniy phisoniy metall graphoniy suspections maked isotops importent becouse wich

I'm not a spelling Nazi but you sure do make it hard to want to read your posts.

Edited by fleabay

Share this post


Link to post
Share on other sites
41 minutes ago, DaTueOwner said:

Yes because realistic light interaction requires simulation of physic laws and therefore graphics is involves physical calculation i.e. solving the redering equation. So I would say you can't really separate those concepts -> realistic looking water needs "100%" physics

It really doesn't. In actual physics, the diffraction of water comes from the movement of light through the whole body of water and the way individual photons interact millions upon millions of times with individual atoms. Similarly the surface of water moves because of the interaction of trillions of molecules.

No game models this or anything close to it. Graphics is the art of imitating physical reality not the art of reproducing it. Describing the way shaders mimic the look of water as "physics" is reducing the term to near meaninglessness.

Share this post


Link to post
Share on other sites

No one write game engines from scratch. 20 years ago too.

You can take some demo's from github:

https://github.com/SDraw/run-on-coal

https://github.com/JoeyDeVries/Cell

http://tesseract.gg/

And try to merge them together.

Becouse every single part of game engine now is:

hmph, it's time to become.... sse/avx/neon hacker, gapi hacker, physics hacker, network hacker and other.

 

Share this post


Link to post
Share on other sites

Well the same can be said about game physics, it's just an approximation of reality. It uses a simplified model (based on classical mechanics?) to describe the interaction of physical objects. Light transportation can be described as lightrays going around in the scene, not including effects like polarisation etc.. However it's a good enough physical model of our visible world. 

I guess what I am trying to express is that everything related to game physics (including optics and classical mechanics) is based on a model, a simplified view of the world that is still able to produce convincing results (to an extend). 

Just now, Makusik Fedakusik said:

No one write game engines from scratch. 20 years ago too.

You can take some demo's from github:

https://github.com/SDraw/run-on-coal

https://github.com/JoeyDeVries/Cell

http://tesseract.gg/

And try to merge them together.

Becouse every single part of game engine now is:

hmph, it's time to become.... sse/avx/neon hacker, gapi hacker, physics hacker, network hacker and other.

 

Guess I'll start with a lighting engine at first, the other parts or not that interesting to me at least. Thanks for the links ;)

Share this post


Link to post
Share on other sites

 

4 hours ago, Irusan, son of Arusan said:

It really doesn't. In actual physics, the diffraction of water comes from the movement of light through the whole body of water and the way individual photons interact millions upon millions of times with individual atoms. Similarly the surface of water moves because of the interaction of trillions of molecules.

No game models this or anything close to it. Graphics is the art of imitating physical reality not the art of reproducing it. Describing the way shaders mimic the look of water as "physics" is reducing the term to near meaninglessness.

By one hand you is absolutely right. Models that fundametal phisics usually utilize  to find fundamental lows,factors (especially factors) and so on really works on micro-level models and simulate trillions of moleculas, atoms and so on. And obviuously it kind of simulation require a month of calculations on super digit-mills per simulation, and to slow and expensive ever not for games, but for applicative sciences and engeneering software too .

But by other hands you is absolutely wrong.  Apllicative sciences and enginering software usualy simulating same processes on macro-level, utilizing fundamental laws and factors found by fundamental science on micro-level (i.e. it integrate behavior of huge clusters or ever whole system, instead of simulation of each indnividual athom ). It enougt to make simulation with accuracy enought for most applicative researches and engeeniting needs, and require a hours or maximum days per simulation on high-end PC or maximum tiny cluster.

But it schemes of integration also have a nonliniar dependence of speed from precission. So optimizing it by speed for cost of accuraacy(or better say by putting it on edge of numerical stability, that makes it much harder to implement ) and limiting some simulation capabilities we have a schemes good for realtime simulations of its effects for games. For example most of people here know classical example of 2D fire simulation that works wery wery fast (at least it able to reproduce fire effect for 320x200 area on 80386@33MHz with high FPS). But it scheme comes from scientific scheme used to model natural convection, just with some limitations like fixed direction of fire that reduce required calculation approximately 6x times and lovest possible (1 byte) precission for field of themperature. On my university years i  used same scheme for scientific simulations. Optimized by accutacy and without limitations it give 0.1 FPS for 100x100 area on Pentium-II@400MHz, and for modelling of flame give results that have no significant visual differences with scheme optimized by speed. Say more modern simulators for 3D fire and smoke utilize same scheme, but mainly using a Lagrange aproach (particles) instead Euler approach (grids) for simulate effect.

An so on with anything else. For example high-end simulation of destruction and deformation (including fabric and so on) use same schemes that CAD software use for strength calculations.

Edited by Fulcrum.013

Share this post


Link to post
Share on other sites
Guest
This topic is now closed to further replies.

  • Advertisement
×

Important Information

By using GameDev.net, you agree to our community Guidelines, Terms of Use, and Privacy Policy.

GameDev.net is your game development community. Create an account for your GameDev Portfolio and participate in the largest developer community in the games industry.

Sign me up!