创建MetalView

    self.mtkView.device = MTLCreateSystemDefaultDevice(); // 获取默认的device
    self.view = self.mtkView;
    self.mtkView.delegate = self;
    self.viewportSize = (vector_uint2){self.mtkView.drawableSize.width, self.mtkView.drawableSize.height}; 

创建pipeline

id<MTLLibrary> defaultLibrary = [self.mtkView.device newDefaultLibrary]; // .metal
    id<MTLFunction> vertexFunction = [defaultLibrary newFunctionWithName:@"vertexShader"]; // 顶点shader，vertexShader是函数名
    id<MTLFunction> fragmentFunction = [defaultLibrary newFunctionWithName:@"samplingShader"]; // 片元shader，samplingShader是函数名
    
    MTLRenderPipelineDescriptor *pipelineStateDescriptor = [[MTLRenderPipelineDescriptor alloc] init];
    pipelineStateDescriptor.vertexFunction = vertexFunction;
    pipelineStateDescriptor.fragmentFunction = fragmentFunction;
    pipelineStateDescriptor.colorAttachments[0].pixelFormat = self.mtkView.colorPixelFormat;
    self.pipelineState = [self.mtkView.device newRenderPipelineStateWithDescriptor:pipelineStateDescriptor
                                                                         error:NULL]; // 创建图形渲染管道，耗性能操作不宜频繁调用
    self.commandQueue = [self.mtkView.device newCommandQueue]; // CommandQueue是渲染指令队列，保证渲染指令有序地提交到GPU

创建顶点矩阵

static const LYVertex quadVertices[] =
    {   // 顶点坐标，分别是x、y、z、w；    纹理坐标，x、y；
        { {  0.5, -0.5, 0.0, 1.0 },  { 1.f, 1.f } },
        { { -0.5, -0.5, 0.0, 1.0 },  { 0.f, 1.f } },
        { { -0.5,  0.5, 0.0, 1.0 },  { 0.f, 0.f } },
        
        { {  0.5, -0.5, 0.0, 1.0 },  { 1.f, 1.f } },
        { { -0.5,  0.5, 0.0, 1.0 },  { 0.f, 0.f } },
        { {  0.5,  0.5, 0.0, 1.0 },  { 1.f, 0.f } },
    };
    self.vertices = [self.mtkView.device newBufferWithBytes:quadVertices
                                                 length:sizeof(quadVertices)
                                                options:MTLResourceStorageModeShared]; // 创建顶点缓存
    self.numVertices = sizeof(quadVertices) / sizeof(LYVertex); // 顶点个数

创建纹理

- (void)setupTexture {
   UIImage *image = [UIImage imageNamed:@"abc"];
   // 纹理描述符
   MTLTextureDescriptor *textureDescriptor = [[MTLTextureDescriptor alloc] init];
   textureDescriptor.pixelFormat = MTLPixelFormatRGBA8Unorm;
   textureDescriptor.width = image.size.width;
   textureDescriptor.height = image.size.height;
   self.texture = [self.mtkView.device newTextureWithDescriptor:textureDescriptor]; // 创建纹理
   
   MTLRegion region = {{ 0, 0, 0 }, {image.size.width, image.size.height, 1}}; // 纹理上传的范围
   Byte *imageBytes = [self loadImage:image];
   if (imageBytes) { // UIImage的数据需要转成二进制才能上传，且不用jpg、png的NSData
       [self.texture replaceRegion:region
                   mipmapLevel:0
                     withBytes:imageBytes
                   bytesPerRow:4 * image.size.width];
       free(imageBytes); // 需要释放资源
       imageBytes = NULL;
   }
}

- (Byte *)loadImage:(UIImage *)image {
   // 1获取图片的CGImageRef
   CGImageRef spriteImage = image.CGImage;
   
   // 2 读取图片的大小
   size_t width = CGImageGetWidth(spriteImage);
   size_t height = CGImageGetHeight(spriteImage);
   
   Byte * spriteData = (Byte *) calloc(width * height * 4, sizeof(Byte)); //rgba共4个byte
   
   CGContextRef spriteContext = CGBitmapContextCreate(spriteData, width, height, 8, width*4,
                                                      CGImageGetColorSpace(spriteImage), kCGImageAlphaPremultipliedLast);
   
   // 3在CGContextRef上绘图
   CGContextDrawImage(spriteContext, CGRectMake(0, 0, width, height), spriteImage);
   
   CGContextRelease(spriteContext);
   
   return spriteData;
}

渲染

// 每次渲染都要单独创建一个CommandBuffer
    id<MTLCommandBuffer> commandBuffer = [self.commandQueue commandBuffer];
    MTLRenderPassDescriptor *renderPassDescriptor = view.currentRenderPassDescriptor;
    // MTLRenderPassDescriptor描述一系列attachments的值，类似GL的FrameBuffer；同时也用来创建MTLRenderCommandEncoder
    if(renderPassDescriptor != nil)
    {
        renderPassDescriptor.colorAttachments[0].clearColor = MTLClearColorMake(0.0, 0.5, 0.5, 1.0f); // 设置默认颜色
        id<MTLRenderCommandEncoder> renderEncoder = [commandBuffer renderCommandEncoderWithDescriptor:renderPassDescriptor]; //编码绘制指令的Encoder
        [renderEncoder setViewport:(MTLViewport){0.0, 0.0, self.viewportSize.x, self.viewportSize.y, -1.0, 1.0 }]; // 设置显示区域
        [renderEncoder setRenderPipelineState:self.pipelineState]; // 设置渲染管道，以保证顶点和片元两个shader会被调用
        
        [renderEncoder setVertexBuffer:self.vertices
                                offset:0
                               atIndex:0]; // 设置顶点缓存

        [renderEncoder setFragmentTexture:self.texture
                                  atIndex:0]; // 设置纹理
        
        [renderEncoder drawPrimitives:MTLPrimitiveTypeTriangle
                          vertexStart:0
                          vertexCount:self.numVertices]; // 绘制
        
        [renderEncoder endEncoding]; // 结束
        
        [commandBuffer presentDrawable:view.currentDrawable]; // 显示
    }
    
    [commandBuffer commit]; // 提交；

名词解释

1. vertex:顶点函数
2. fragment: 片元函数
3. buffer: vertex参数,小括号模具代表了encoder设置VertexBuffer的index，
4. texture: 输入纹理，小括号枚举代表了encoder设置的texture的index

疑问点

1. sharder是怎么绑定的