Metal: Compute Shaders (#19)

* check for too bix texture bindings

* implement lod query

* print shader stage name

* always have fragment input

* resolve merge conflicts

* fix: lod query

* fix: casting texture coords

* support non-array memories

* use structure types for buffers

* implement compute pipeline cache

* compute dispatch

* improve error message

* rebind compute state

* bind compute textures

* pass local size as an argument to dispatch

* implement texture buffers

* hack: change vertex index to vertex id

* pass support buffer as an argument to every function

* return at the end of function

* fix: certain missing compute bindings

* implement texture base

* improve texture binding system

* remove useless exception

* move texture handle to texture base

* fix: segfault when using disposed textures

---------

Co-authored-by: Samuliak <samuliak77@gmail.com>
Co-authored-by: SamoZ256 <96914946+SamoZ256@users.noreply.github.com>

src/Ryujinx.Graphics.GAL/Multithreading/ThreadedPipeline.cs

@@ -63,9 +63,9 @@ namespace Ryujinx.Graphics.GAL.Multithreading
             _renderer.QueueCommand();
         }
 
-        public void DispatchCompute(int groupsX, int groupsY, int groupsZ)
+        public void DispatchCompute(int groupsX, int groupsY, int groupsZ, int groupSizeX, int groupSizeY, int groupSizeZ)
         {
-            _renderer.New<DispatchComputeCommand>().Set(groupsX, groupsY, groupsZ);
+            _renderer.New<DispatchComputeCommand>().Set(groupsX, groupsY, groupsZ, groupSizeX, groupSizeY, groupSizeZ);
             _renderer.QueueCommand();
         }