What Lanquage "Phun/Algodoo" made with?

[Rhett97]

Basically the Title,Because iv been wanting to make a game and i want to know WHAT Language and Programs were used....And how the Phuck you made your own Lanquage for it.

[standardtoaster]

It was made with C++ but I don't know how they made thyme.

[admin]

Yes, it is all coded in C++ (with some things closer to C) and Thyme itself is also defined and implemented in C++. The graphics is done in OpenGL. Some standard libraries like STL and BOOST are also used. Future releases will also use OpenCL for parallel simulations on GPU and multi-core.

[Jaxc]

swedish

[Chronos]

swedish

you, sir, have just won

[Jallaboy]

swedish

you, sir, have just won

lal, thoght of the same as jaxc, but he was faster

[Rhett97]

swedish

you, sir, have just won

lal, thoght of the same as jaxc, but he was faster

yeah about a week faster

[macunixgeek]

OpenCL for parallel simulations on GPU

My brain just exploded. THE DAY THAT GPU COMPUTATION IS ADDED TO ALGODOO IS THE DAY THAT WE CAN SIMULATE 100000 BODIES WITHOUT LAG!!!

[Sniperkasa]

OMG thats true by then I just got to fix my issue with the water.. water works fine but I get some hardware water generation error every once in a while

[admin]

I can gossip some...

We have water-simulations running on GPU. We can do a million water particles in 3D at 2Hz and these fluids are nearly perfectly incompressible and fully integrated with rigid body simulations. I reckon we will be able to do 200.000 particles at fully interactive frame rates in Algodoo. We'll soon be releasing some really cool videos of our 3D fluids.

Simulating rigid body contacts on GPU is harder since the equations are inequalities (for example, Ax=b where x>0) and not just equalities (Ax = b). x here refers to the contact force and for large systems there can be tens of thousands of these forces that need to be solved for. Inequalities require much more communication in the processor, and communication is something a GPU is really bad at. In fact, it is pretty much an open research problem how to do this well on GPU, but we have our ways...

[immibis]

OpenCL for parallel simulations on GPU

My brain just exploded. THE DAY THAT GPU COMPUTATION IS ADDED TO ALGODOO IS THE DAY THAT WE CAN SIMULATE 100000 BODIES WITHOUT LAG!!!

Only for those of us lucky enough to have good GPUs. Algodoo should split the load between the GPU and CPUs/cores to reduce lag as much as possible. That would be a neat feature.

[Sonic]

That would be awesome. Finally I could make my water park that I've been dreaming about since the very first day I got Phun! May that day come soon!!!

[thebeech]

Only for those of us lucky enough to have good GPUs. Algodoo should split the load between the GPU and CPUs/cores to reduce lag as much as possible. That would be a neat feature.

didnt ken just say that gpu is slow for communication with other devices ?

Inequalities require much more communication in the processor, and communication is something a GPU is really bad at.

[Sonic]

Oh...so my dream once again has to wait...

[macunixgeek]

Couldn't the GPU just get a batch of inequalities (one for each core) then have each core solve one, and transmit the entire batch of solutions back to the CPU? That would minimize communications, especially if the solution data was compressed.

[Tank2333]

Only for those of us lucky enough to have good GPUs. Algodoo should split the load between the GPU and CPUs/cores to reduce lag as much as possible. That would be a neat feature.

didnt ken just say that gpu is slow for communication with other devices ?

Inequalities require much more communication in the processor, and communication is something a GPU is really bad at.

he said that the gpu is has bad communication in the gpu itself
a gpu has much streaming processors wich can calculate simultanly
and those streaming processers are like cooks in a canteen kitchen if there are 100 cooks and 100 meals are have to be made 1 cook makes on meal and all is fine but when only 1 meal is requiered to make than 100 cooks cooking at 1 meal and that is not very effecient
so a gpu is good at proceeding single tasks simultan but
so i think the problem is that water has only 1 input and thats easy for gpu
but a rigid body has more than 1 force to be calculated like rotaion friction
im no programmer i cant say that for sure
but i think that the GPU has big future in physical programms

does that makes sense anyway?... i hope u get what i try to explain ^^

[Altro]

OpenCL for parallel simulations on GPU

My brain just exploded. THE DAY THAT GPU COMPUTATION IS ADDED TO ALGODOO IS THE DAY THAT WE CAN SIMULATE 100000 BODIES WITHOUT LAG!!!

Only for those of us lucky enough to have good GPUs. Algodoo should split the load between the GPU and CPUs/cores to reduce lag as much as possible. That would be a neat failure.

[izacque]

I can gossip some...

We have water-simulations running on GPU. We can do a million water particles in 3D at 2Hz and these fluids are nearly perfectly incompressible and fully integrated with rigid body simulations. I reckon we will be able to do 200.000 particles at fully interactive frame rates in Algodoo. We'll soon be releasing some really cool videos of our 3D fluids.

Simulating rigid body contacts on GPU is harder since the equations are inequalities (for example, Ax=b where x>0) and not just equalities (Ax = b). x here refers to the contact force and for large systems there can be tens of thousands of these forces that need to be solved for. Inequalities require much more communication in the processor, and communication is something a GPU is really bad at. In fact, it is pretty much an open research problem how to do this well on GPU, but we have our ways...

3D fluids!??!?!!? are you hinting at a 3D physics engine?

[KarateBrot]

3D fluids!??!?!!? are you hinting at a 3D physics engine?

algoryx is more than just algodoo they make 3d physics simulations as well
http://www.algoryx.se/products.html

[izacque]

dang... that is some good stuff you got there, algoryx