Jump to content

  • Log In with Google      Sign In   
  • Create Account

Promit

Member Since 29 Jul 2001
Offline Last Active Today, 01:02 PM

#5307731 So, I want to become a game devolper...

Posted by on 24 August 2016 - 05:40 PM

I like C# a lot, you can grab Visual Studio community and immediately start building apps that run on Android, iOS, and Windows (phone/desktop/Xbox). Really can't ask for much more.

Unity3D is a solid choice for games, but it won't get you much in terms of building non-game applications.




#5305130 Chosing An Engine For A Heavy Physics-Based Game

Posted by on 10 August 2016 - 09:45 AM

We specialize in physics-based character animations, and I built our own engine to underpin everything. That was a choice made for a variety of reasons, but the big one for this discussion is that our physics is a customized version of Bullet. The stock library isn't sufficiently stable in solving certain types of problems (joints, soft body interactions) and so we went in and made the necessary changes ourselves. I don't think that's an ability any major engine gives you these days. Everyone assumes that oh, well we can just use the raw high level features in PhysX/Havok and that's enough to build what we need. It wasn't the case in doing physics animation, for us.

 

That said, I'd probably just roll with Unreal these days unless you have the experience and manpower to build from nothing.




#5304281 Cast your vote now!

Posted by on 05 August 2016 - 06:44 PM

I'm real concerned about why Notepad++ is in the list.




#5304093 How Do You Test/check If Lookat Matrix Is Correct?

Posted by on 04 August 2016 - 07:01 PM

If you invert the matrix, you'll get three rows with each of the three basis axes of the camera (xaxis, yaxis, zaxis as labeled in your code) and the cameraPosition in the fourth row.




#5304074 Should I Get A Low-Level Or A High-Level Lib For A Really Basic, Not Secured,...

Posted by on 04 August 2016 - 04:53 PM

 

why can't just people use simple words


Because distributed systems and networking is, unfortunately, not a simple problem.

 

That said, you don't need a lot of the things they're talking about to start out. Look at the chat example in the RakNet source. It's enough to get a LAN game working.




#5303840 Should I Get A Low-Level Or A High-Level Lib For A Really Basic, Not Secured,...

Posted by on 03 August 2016 - 12:23 PM

RakNet is still by far the easiest system to get up and running, do cool things with, and ship with. ENet is a simpler tool worth looking at.




#5302945 How Do You Deal With Errors On Gpus? Do You At All?

Posted by on 28 July 2016 - 10:01 AM

I'm in the camp that yes, (consumer) GPUs are not as reliable as CPUs. They take a lot of shortcuts and optimizations that generate nearly-correct results. That said, a blue screen failure due to a specific operation is definitely out of the ordinary, though using Vulkan in these early days definitely exposes you to it more than it otherwise would. This actually does sound like you should build a test case and submit it to AMD, as a lot of these weird corner cases can show up dependent on the GPU and they may not have noticed.

 

As far as handling errors - it is common to have workarounds for certain hardware configurations that are known to break. It is also common to have workarounds for particular drivers, particular operating systems, etc. These are all derived from testing before release (or afterwards...) to find out what works and what doesn't. What you don't have, though, is the ability to detect and handle GPU or driver errors in any sensible way. A blue screen is a kernel mode unhandled exception, and there is not a damn thing you can do about it after the bug has been invoked.




#5302705 you all say oculus rift but why not google glass?

Posted by on 26 July 2016 - 09:53 PM

I have a HoloLens. It's a really cool toy, but attempting to sell this device to the general public would be a disaster and MS knows it. I think there's a wonderful future ahead for AR googles when they get the tech and use cases nailed, but VR is here now and AR is just a prototype.

 

By the way, the surprising amazing thing about the HoloLens isn't the holograms - it's the world tracking. I can lay down holograms in a couple offices and the hallways, and wander about freely and it always knows exactly where I am. And it persists the world correctly at all times. No cameras or laser projectors or outboard sensors needed. This makes room-scale VR look like child's play. 

 

Lastly, on the subject of VR: Vive is fantastic. The dual tracked controllers actually give you a real tactile sense into the world. Oculus, on the other hand, sucks due to not having it. (Although their optics might be better, I'm still deciding.) Being in VR and using an Xbox controller to play is weird and disconnected and fundamentally not that fun. (The single camera setup also means Oculus has some nasty occlusion issues.) Vive feels like what VR should be. Hopefully Oculus Touch will address the issue later this year.




#5300389 Posibility of getting to game industry?

Posted by on 12 July 2016 - 09:09 AM

It wouldn't be devastating for a AAA company to invest a month in allowing a developer to learn C++ - the pay would probably reflect it. The lead physics programmer I work with was hired with zero C++ experience (early 2000s). I'm not going to say this is the case now but it's certainly happened in the past.

A month? A month of C++ experience is just enough to really get in the groove of breaking everything. You'll have to do detailed code reviews on every check-in for a good while. The early 2000s were a different time - industry C++ code was much closer to C, lots of people had C experience, and frankly the industry was small and much easier to get into. Each successive generation of tech has brought increased requirements for candidates as games themselves have evolved into extremely sophisticated systems.

 

There are a few exceptions for people who are not being hired for their programming prowess. Physics people were rare and massively in demand for several years from about 2002 - this was right when Havok was becoming a big name and HL2 was about to arrive. Even now they're in demand, but the good physics people all know C++ now  :rolleyes: Same thing happened when the PS3 arrived for low level optimization/SPU people, high end GPGPU/shader people, etc. C++ was not required for those jobs because they were based on specialized hard-to-find experience.

 

maths/physics

 
This is certainly the skill that I really am bad at.

 

That is something you will need to fix, no exceptions.




#5299844 Posibility of getting to game industry?

Posted by on 08 July 2016 - 10:31 PM

 

 

is a great, productive language to get a lot of things done in. But C++ gets you the jobs, not C#.

 
HI, just wanna know if this is kind of true? like for example the industry prefers c++ people?

 


Promit knows whereof he speaks.

 

When it comes to the game industry, it's simply an expected and generally required part of your skill set. It's been the standard for a very long time now, and there's a reasonably healthy supply of capable candidates with that feather in their cap. If someone comes in with solid C++ abilities and not too much else, it's assumed they'll pick up other things as needed. Whether that's justified or not, that is an intrinsic widespread attitude in the industry. It's never assumed that a junior hire will just pick up C++, so the company basically plans to have to train you for a while before you're capable. And if I'm choosing between the candidate I have to train or the candidate who is ready to go out of the gate, well... it can be overcome, but it's easier just to have that knowledge up front.

 

This comment is ONLY applicable to the game industry. There are many types of software development that don't have this expectation.




#5299400 Is DirectXMath thread safe?

Posted by on 06 July 2016 - 04:58 PM

Don't all of the DirectXMath functions operate on local by-value parameters and return by value? They don't know anything about threads or references, they're just functions operating on data. Anything YOU do with assigning or reading thread-shared variables is on you, and no different from anything else.




#5299352 Are there too many Unity Developers?

Posted by on 06 July 2016 - 11:34 AM

At your stage in the game, producing something is vastly more important than what you use to make it. Use whatever gets you to the goal of a completed project.

 

Long term, "Unity" or "CryEngine" or "Unreal" are not generally sufficient to get a job without a portfolio of projects AND significant primary skills, either art or programming.




#5299328 Posibility of getting to game industry?

Posted by on 06 July 2016 - 09:22 AM

2nd. As a person wanting to get into game industry, Is it right to aim for bigger game dev industry? like EA or SE?

Sure. One of the positive things about the game industry is that it's wonderfully neutral to your upbringing, background, educational background, etc. If you can do the job, you'll probably get the job. I feel a little bit like economic downturn in recent years has made it a bit harder than it used to be, but it's absolutely doable.

Will continuing c++ be a good asset for me? or Just use c# which a lot of inie game dev use?
C# is a great, productive language to get a lot of things done in. But C++ gets you the jobs, not C#.

3rd. What are the chances that at my mid 20s I can be able to get into game industry? How high would you think that be?
If you can learn enough to keep up, then you'll get a job. It'll be entry level for sure, and it will be challenging, but it's not an unrealistic goal. 


#5298917 Porting OpenGL to Direct3D 11 : How to handle Input Layouts?

Posted by on 03 July 2016 - 01:57 PM

Won't claim this is the best/only way of doing this, but I define a "Geometry Input" object that is more or less equivalent to a VAO. It holds a vertex format and the buffers that are bound all together in one bundle. The vertex format is defined identically to D3D11_INPUT_ELEMENT_DESC in an array. In GL, this pretty much just maps onto a VAO. (It also virtualizes neatly to devices that don't have working implementations of VAO. Sadly they do exist.) In D3D, it holds an input layout plus a bunch of buffer references and the metadata for how they're bound to the pipeline. This is as simple as can be:

ID3D11InputLayout* _layout;
std::vector<ID3D11Buffer*> _vertexBuffers;
std::vector<UINT> _strides;
std::vector<UINT> _offsets;

That's more or less sufficient to emulate VAO for the base case. You can create duplicate input layouts for each input, or build a hashed cache and look them up. There's one nasty catch, which is the shader bytecode. I decided to make it optional to supply a shader to create one of these objects (but mandatory to supply the vertex format array). If you do supply a shader, the GL path ignores it and the D3D path uses the stored bytecode. If you don't supply it, I actually generate a fake shader matching the supplied vertex format, and compile a bytecode from it. This sounds annoying but is actually really easy to do. 




#5298707 Creating OpenGL context in different OpenGL driver versions

Posted by on 01 July 2016 - 10:33 AM

So for example specify 4.5 version in profile struct but driver responses with 4.1 version then this is what you could get.

That's not legal behavior. The spec is very clear that the implementation is not allowed to respond with a version less than requested in the attribs submitted to wglCreateContextAttribsARB.
 
 

Same is true for if you want a core profile where depricated functions got out of scope or compatibility one where you still could use glBegin, glEnd ...

If you ask for a core profile, the implementation is required to give you a core profile. Incidentally, a core profile has a slight negative performance impact on NVIDIA and probably on the other major manufacturers too. It's not recommended to use it unless you have a concrete reason.
 

The reason is that if you specify a specific version driver is going to load that version function code into process memory and then you could access that versions extensions first.

No driver I've ever seen works that way. The driver is a single monolithic implementation which supports some particular version of OpenGL. It will load in its entirety and respond with that version of OpenGL no matter what you request during context creation. (Mac might be a bit different since it only has a few versions, I can't remember now.) It wouldn't be practical to slice the versions into separate libraries and defer loads, as that would create more problems without any benefits.






PARTNERS