Destroy vs Release render texture

[j04milan]

Hi i’m creating an app in which I’m creating and destroying render textures based on different images sizes but I don’t know exactly what method use for destroying a render texture before creating another one. Should I use Destroy or Release the render texture to create another one with different size. What’s the difference?? Thanks.

[arkano22]

@j04milan

Hi i’m creating an app in which I’m creating and destroying render textures based on different images sizes but I don’t know exactly what method use for destroying a render texture before creating another one. Should I use Destroy or Release the render texture to create another one with different size. What’s the difference?? Thanks.

The documentation explains it pretty well:
https://docs.unity3d.com/ScriptReference/RenderTexture.Release.html

Releases the RenderTexture.

This function releases the hardware resources used by the render texture. The texture itself is not destroyed, and will be automatically created again when being used.

In layman terms, Release() destroys the GPU (aka “hardware”) resources associated to the render texture, but the RenderTexture object used by Unity to wrap around this GPU data and expose it nicely to you is kept around and will re-generate the underlying GPU data if used again, as it would if you manually called Create().

Destroy() however, releases both GPU memory and the RenderTexture object.

So which one to use depends on the lifecycle of your render textures, if you’re reusing them in a pool-like fashion, etc.

The documentation for RenderTexture itself goes into further detail:

A render texture only has a data representation on the GPU and you need to use Texture2D.ReadPixels to transfer its contents to CPU memory.

Keep in mind that render texture contents can become “lost” on certain events, like loading a new level, system going to a screensaver mode, in and out of fullscreen and so on. When that happens, your existing render textures will become “not yet created” again, you can check for that with IsCreated function.

As with other “native engine object” types, it is important to pay attention to the lifetime of any render textures and release them when you are finished using them with the Release function, as they will not be garbage collected like normal managed types.

[j04milan]

@arkano22

The documentation explains it pretty well:
https://docs.unity3d.com/ScriptReference/RenderTexture.Release.html

In layman terms, Release() destroys the GPU (aka “hardware”) resources associated to the render texture, but the RenderTexture object used by Unity to wrap around this GPU data and expose it nicely to you is kept around and will re-generate the underlying GPU data if used again, as it would if you manually called Create().

Destroy() however, releases both GPU memory and the RenderTexture object.

So which one to use depends on the lifecycle of your render textures, if you’re reusing them in a pool-like fashion, etc.

The documentation for RenderTexture itself goes into further detail:

Thanks for the explanation.there is something I don’t quite understan, since you have to use new Rendertexture to change the size of the rendertexture your making reference. wouldn’t it pile up if I use only Release () instead of Destroy?. I’m using a Rendertexture that changes its size everytime y pick an image so in order to change its size I have to use new Rendertexture (). will it create multiple instances of the rendertexture if I use only release () instead of Destroy() ?? Thanks

[arkano22]

@j04milan

Thanks for the explanation.there is something I don’t quite understan, since you have to use new Rendertexture to change the size of the rendertexture your making reference.

That doesn’t change the size of the texture, it creates an entirely new texture of a different size. Anytime you use the “new” keyword in any programming language, it means you’re creating a new object (and allocating memory for it).

To resize a texture, you Release() its data and then change its size:

rt.Release();
rt.width = newWidth;
rt.height = newHeight;
rt.Create();

@j04milan

wouldn’t it pile up if I use only Release () instead of Destroy?. I’m using a Rendertexture that changes its size everytime y pick an image so in order to change its size I have to use new Rendertexture (). will it create multiple instances of the rendertexture if I use only release () instead of Destroy() ?? Thanks

Yep of course, if you create a new RenderTexture and only Release() it but never call Destroy() then the GPU data will be released but the RenderTexture object will be kept around forever.

Rule of thumb:
to “undo” rt = new RenderTexture() call Destroy(rt).
to “undo” rt.Create() call rt.Release().

kind regards,

[j04milan]

@arkano22

That doesn’t change the size of the texture, it creates an entirely new texture of a different size. Anytime you use the “new” keyword in any programming language, it means you’re creating a new object (and allocating memory for it).

To resize a texture, you Release() its data and then change its size:

rt.Release();
rt.width = newWidth;
rt.height = newHeight;
rt.Create();

Rule of thumb:
to “undo” rt = new RenderTexture() call Destroy(rt).
to “undo” rt.Create() call rt.Release().

kind regards,

Thanks bro just the explanation I needed.