Community Hardware Survey

[Racers52]

This poll should generate a community census on the hardware being used to play Mx Simulator.

Q.1 Having trouble finding your PC specifications?

Find basic system information in Settings
You can see similar information (with some added details about the version of Windows you are running) in Settings.

1. Click the Start button and choose Settings (it's the gear-shaped icon above the power icon).

2. Click "System."

3. Click "About."



Q.2 Having trouble finding what graphics card you have?

1. Click Start.

2. On the Start menu, click Run.

3. In the Open box, type "dxdiag" (without the quotation marks), and then click OK.

4. The DirectX Diagnostic Tool opens. Click the Display tab.

5. On the Display tab, information about your graphics card is shown in the Device section. You can see the name of your card, as well as how much video memory it has.

[mxer433]

could it be

[Wahlamt]

I think in general, if your pc isn't even good enough to launch a AAA title, you can't expect to run any game good.

[Racers52]

I think in general, if your pc isn't even good enough to launch a AAA title, you can't expect to run any game good.

The minimum requirements for AAA titles are pretty beefy nowadays, just goes to show people have better systems than what you would expect.

Hoping more people vote so there is a better understanding of the hardware people are running in this community!

[jlv]

I think in general, if your pc isn't even good enough to launch a AAA title, you can't expect to run any game good.

A lot of people got started on this game using terrible integrated graphics that couldn't launch a AAA game. I made a lot of sales by supporting those systems. With that said, it's a pain to support both the OpenGL 1.x fixed function pipeline and the OpenGL 2.x shader pipeline. Also, GPUs that only support 1.x are pretty much non-existent at this point. So I'm planning on leaving 1.x behind soon. If you're on a 10 year old craptop it's going to be time to upgrade.

[nascarnate]

I think in general, if your pc isn't even good enough to launch a AAA title, you can't expect to run any game good.

A lot of people got started on this game using terrible integrated graphics that couldn't launch a AAA game. I made a lot of sales by supporting those systems. With that said, it's a pain to support both the OpenGL 1.x fixed function pipeline and the OpenGL 2.x shader pipeline. Also, GPUs that only support 1.x are pretty much non-existent at this point. So I'm planning on leaving 1.x behind soon. If you're on a 10 year old craptop it's going to be time to upgrade.

Would this mean only opengl 2.x or moving to the latest and greatest at opengl 4.x?

[Racers52]

Nate asking the important questions

[jlv]

I think in general, if your pc isn't even good enough to launch a AAA title, you can't expect to run any game good.

A lot of people got started on this game using terrible integrated graphics that couldn't launch a AAA game. I made a lot of sales by supporting those systems. With that said, it's a pain to support both the OpenGL 1.x fixed function pipeline and the OpenGL 2.x shader pipeline. Also, GPUs that only support 1.x are pretty much non-existent at this point. So I'm planning on leaving 1.x behind soon. If you're on a 10 year old craptop it's going to be time to upgrade.

Would this mean only opengl 2.x or moving to the latest and greatest at opengl 4.x?

3.x actually removed a lot of stuff. Not sure what they were smoking when they did that. 4.x added tesselation shaders which would be useful for the terrain but it isn't that important. Really, just being able to use pixel and fragment shaders without needing a fixed function fallback would be a huge relief.

[Racers52]

3.x actually removed a lot of stuff. Not sure what they were smoking when they did that. 4.x added tesselation shaders which would be useful for the terrain but it isn't that important. Really, just being able to use pixel and fragment shaders without needing a fixed function fallback would be a huge relief.

So we are on a 2.x version now and we would be moving to 4.x correct?

[jlv]

So we are on a 2.x version now and we would be moving to 4.x correct?

It's kind of fuzzy with OpenGL. There are lots of extensions that are commonly used before they're standardized into the official spec. For MXS, it requires 1.1. It can optionally use multitexture (standardized in 1.3), vertex and fragment shaders (2.0), and framebuffer objects (3.0).

So when I say leave 1.x behind I mean it will drop the fixed function / no shaders rendering path. I'd also stop trying to deal with small maximum texture sizes which isn't technically a 1.x thing but the GPUs with the small texture limits are the same ones.

[Racers52]

So we are on a 2.x version now and we would be moving to 4.x correct?

It's kind of fuzzy with OpenGL. There are lots of extensions that are commonly used before they're standardized into the official spec. For MXS, it requires 1.1. It can optionally use multitexture (standardized in 1.3), vertex and fragment shaders (2.0), and framebuffer objects (3.0).

So when I say leave 1.x behind I mean it will drop the fixed function / no shaders rendering path. I'd also stop trying to deal with small maximum texture sizes which isn't technically a 1.x thing but the GPUs with the small texture limits are the same ones.

So what base version would we be moving to? How much code writing/re-writing do you anticipate?

[jlv]

So we are on a 2.x version now and we would be moving to 4.x correct?

It's kind of fuzzy with OpenGL. There are lots of extensions that are commonly used before they're standardized into the official spec. For MXS, it requires 1.1. It can optionally use multitexture (standardized in 1.3), vertex and fragment shaders (2.0), and framebuffer objects (3.0).

So when I say leave 1.x behind I mean it will drop the fixed function / no shaders rendering path. I'd also stop trying to deal with small maximum texture sizes which isn't technically a 1.x thing but the GPUs with the small texture limits are the same ones.

So what base version would we be moving to? How much code writing/re-writing do you anticipate?

I don't think you understand. OpenGL doesn't completely redo the API on every version. They just add new features to the existing API. (And unfortunately with 3.1 remove features.) I don't have to convert it to 2.x or anything like that. I'd just be requiring features that were standardized in 2.0.

The benefit for requiring shaders is that it would allow me to do the terrain in a more efficient way when I don't have to support the old fixed function pipeline. Basically what I'd be doing is switching from having the CPU calculate a close to ideal terrain mesh from the current viewpoint and sending it to the GPU on each frame to keeping the whole field on the GPU and rendering it in chunks at the highest detail level required for each chunk. So more triangles overall but much less data sent to the GPU.

[Racers52]

I don't think you understand. OpenGL doesn't completely redo the API on every version. They just add new features to the existing API. (And unfortunately with 3.1 remove features.) I don't have to convert it to 2.x or anything like that. I'd just be requiring features that were standardized in 2.0.

The benefit for requiring shaders is that it would allow me to do the terrain in a more efficient way when I don't have to support the old fixed function pipeline. Basically what I'd be doing is switching from having the CPU calculate a close to ideal terrain mesh from the current viewpoint and sending it to the GPU on each frame to keeping the whole field on the GPU and rendering it in chunks at the highest detail level required for each chunk. So more triangles overall but much less data sent to the GPU.

That makes more sense, I thought jumping to different versions would required alot of new work.

I went back into an old topic and found some old points you were talking about.

The 16k texture might have been the total size of the texture atlas. It'd be ridiculous to use that for a single texture as it'd be mapped to something like 1000 DPI assuming the gun has around a square foot of surface area.

I think you'll find it's even more critical to have efficient models and textures when the engine is doing all sorts of fancy multi-pass effects. Do something like dynamic cube environment maps and suddenly you're redrawing the scene 6 times for each shiny object. That makes it critical to keep the geometry under control. The models in the demo obviously have multiple LODs as you can see the LOD popping as the camera moves around. I'm sure the textures are also all very carefully mapped to get the most out of every byte of texture memory.

I don't really want to bother with the pretty/slow stuff since it's hard to do and MXS is more of an esports type of game anyway. The engine changes I'd like to make are all things that would either break even or improve the frame rate.

- Lighting entirely with HDR environment maps. Should be as fast as the current fragment shader.
- Make the terrain tessellator work on larger areas instead of tessellating down to 3x3 triangle fans. It'd make more triangles overall but require much less data to be sent to the GPU. This is one of the biggest performance problems in the current engine.
- Atlas the billboards and render them all in one draw call. This would be much faster but at the cost of perfect sorting for translucent billboards.
- Require large texture support. The current terrain engine goes through awful contortions to stay under 1024x1024 textures. Probably been over 10 years since a card had that limit.
- I'd kind of like to have displacement textures even though that would hurt the FPS.

Is anyone still using anything less than OpenGL 2.0? I should do a poll on that.

The participation rate isnt as large as I would have liked it to be but hopefully the survey helped.

Has anything changed in the last year or are these points still accurate as to what you want to add to the game?

Also for anyone wondering im referencing this topic.

[jlv]

Has anything changed in the last year or are these points still accurate as to what you want to add to the game?

No. That's pretty much exactly what I need to do.