Metalを使用して、シェーダーを使った、最小限のサンプル、覚書
■Step1：初期化

以前の初期化を参照してください
・iOsでObjective-cを使ってMetalを初期化
https://qiita.com/sanoh/items/f123306bd347764fcac4
・MacでObjective-cを使ってMetalを初期化
https://qiita.com/sanoh/items/fa49b9888e24fc977e0d
■Setp2：シェーダー作成

まず「Shader.metal」を作成します。

Shader.metal

#include <metal_stdlib>
using namespace metal;

vertex float4 myVertexShader(const device packed_float3* vertex_array [[ buffer(0) ]],
                             unsigned int vid [[ vertex_id ]]) {
    return float4(vertex_array[vid], 1.0);
}

fragment half4 myFragmentShader() {
    return half4(1.0);
}

■Setp3：MetalRenderer.mの編集

MetalRenderer.m

#import "MetalRenderer.h"

static const NSUInteger MaxBuffersInFlight = 3;

@implementation MetalRenderer
{
    dispatch_semaphore_t _in_flight_semaphore;
    id <MTLDevice> _device;
    id <MTLCommandQueue> _command_queue;
    
    id <MTLRenderPipelineState> _pipeline_state; //＜新規＞
    id <MTLBuffer>   _buffer_position;           //＜新規＞
}

-(nonnull instancetype)initWithMetalKitView:(nonnull MTKView *)view;
{
    //頂点データ＜新規＞
    float vertices_array[]={
        0.0, 1.0, 0.0,
        -1.0, -1.0, 0.0,
        1.0, -1.0, 0.0,
    };
    if(self= [super init])
    {
        _device = view.device;
        _in_flight_semaphore = dispatch_semaphore_create(MaxBuffersInFlight);
        view.colorPixelFormat = MTLPixelFormatBGRA8Unorm;
        _command_queue = [_device newCommandQueue];
        //頂点情報の登録＜新規＞
        NSInteger size = sizeof(vertices_array);
        _buffer_position=[_device newBufferWithBytes:vertices_array 
                                              length:size
                                             options:MTLResourceOptionCPUCacheModeDefault];
        //シェーダーの登録＜新規＞
        id <MTLLibrary> library = [_device newDefaultLibrary];
        MTLRenderPipelineDescriptor *descriptor = [[MTLRenderPipelineDescriptor alloc] init];
        descriptor.vertexFunction = [library newFunctionWithName:@"myVertexShader"];
        descriptor.fragmentFunction = [library newFunctionWithName:@"myFragmentShader"];
        descriptor.colorAttachments[0].pixelFormat=MTLPixelFormatBGRA8Unorm;       
        NSError *error = NULL;
        _pipeline_state = [_device newRenderPipelineStateWithDescriptor:descriptor 
                                                                  error:&error];
        if (!_pipeline_state)
        {
            NSLog(@"Failed to created pipeline state, error %@", error);
        }
    }
    return self;
}

- (void)drawInMTKView:(nonnull MTKView *)view
{
    dispatch_semaphore_wait(_in_flight_semaphore, DISPATCH_TIME_FOREVER);
    
    id <MTLCommandBuffer> command_buffer = [_command_queue commandBuffer];
    __block dispatch_semaphore_t block_sema = _in_flight_semaphore;
    [command_buffer addCompletedHandler:^(id<MTLCommandBuffer> buffer)
     {
         dispatch_semaphore_signal(block_sema);
     }];
    MTLRenderPassDescriptor* render_pass_descriptor = view.currentRenderPassDescriptor;
    render_pass_descriptor.colorAttachments[0].loadAction=MTLLoadActionClear;
    render_pass_descriptor.colorAttachments[0].storeAction=MTLStoreActionStore;
    render_pass_descriptor.colorAttachments[0].clearColor=MTLClearColorMake(1.0, 0.0, 0.0, 1.0);
    if(render_pass_descriptor != nil)
    {
        id <MTLRenderCommandEncoder> render_encoder = [command_buffer renderCommandEncoderWithDescriptor:render_pass_descriptor];
        //モデルの設定＜新規＞
        [render_encoder setRenderPipelineState:_pipeline_state];
        [render_encoder setVertexBuffer:_buffer_position 
                                 offset:0 
                                atIndex:0];
        [render_encoder drawPrimitives:MTLPrimitiveTypeTriangle 
                           vertexStart:0 
                           vertexCount:3];
        
        [render_encoder endEncoding];
        [command_buffer presentDrawable:view.currentDrawable];
    }
    [command_buffer commit];
}

- (void)mtkView:(nonnull MTKView *)view drawableSizeWillChange:(CGSize)size
{
}

@end

■Setp4：実行

iOsは実機で実行してくださいMacの場合は以下のようになります。