Question Custom Effect Pass vs Built In OnRender absolutely tanks performance

I am trying to implement a custom Effect Pass that works in BuiltIn

Code (CSharp):

1.  void OnRenderImage(RenderTexture source, RenderTexture destination) {
2.         kuwaharaMat.SetInt("_KernelSize", kernelSize);
3.         kuwaharaMat.SetInt("_N", 8);
4.         kuwaharaMat.SetFloat("_Q", sharpness);
5.         kuwaharaMat.SetFloat("_Hardness", hardness);
6.         kuwaharaMat.SetFloat("_Alpha", alpha);
7.         kuwaharaMat.SetFloat("_ZeroCrossing", zeroCrossing);
8.         kuwaharaMat.SetFloat("_Zeta", useZeta ? zeta : 2.0f / 2.0f / (kernelSize / 2.0f));
9.  
10.         var structureTensor = RenderTexture.GetTemporary(source.width, source.height, 0, source.format);
11.         Graphics.Blit(source, structureTensor, kuwaharaMat, 0);
12.         var eigenvectors1 = RenderTexture.GetTemporary(source.width, source.height, 0, source.format);
13.         Graphics.Blit(structureTensor, eigenvectors1, kuwaharaMat, 1);
14.         var eigenvectors2 = RenderTexture.GetTemporary(source.width, source.height, 0, source.format);
15.         Graphics.Blit(eigenvectors1, eigenvectors2, kuwaharaMat, 2);
16.         kuwaharaMat.SetTexture("_TFM", eigenvectors2);
17.  
18.         RenderTexture[] kuwaharaPasses = new RenderTexture[passes];
19.  
20.         for (int i = 0; i < passes; ++i) {
21.             kuwaharaPasses[i] = RenderTexture.GetTemporary(source.width, source.height, 0, source.format);
22.         }
23.  
24.         Graphics.Blit(source, kuwaharaPasses[0], kuwaharaMat, 3);
25.  
26.         for (int i = 1; i < passes; ++i) {
27.             Graphics.Blit(kuwaharaPasses[i - 1], kuwaharaPasses[i], kuwaharaMat, 3);
28.         }
29.  
30.         //Graphics.Blit(structureTensor, destination);
31.         Graphics.Blit(kuwaharaPasses[passes - 1], destination);
32.  
33.         RenderTexture.ReleaseTemporary(structureTensor);
34.         RenderTexture.ReleaseTemporary(eigenvectors1);
35.         RenderTexture.ReleaseTemporary(eigenvectors2);
36.         for (int i = 0; i < passes; ++i) {
37.             RenderTexture.ReleaseTemporary(kuwaharaPasses[i]);
38.         }
39.     }

and this is my attempt at turning in in a custom Feature with Custom pass:

Code (CSharp):

1.  class KuawaharaRenderPass : ScriptableRenderPass
2.     {
3.         private ProfilingSampler m_ProfilingSampler = new ProfilingSampler("KuawaharaPass");
4.         private static Material _material;
5.  
6.  
7.         private bool isUsingA = false;
8.  
9.         private RTHandle _swapBufferA;
10.         private RTHandle _swapBufferB;
11.         private RTHandle _TFMBuffer;
12.  
13.         public RTHandle NextBuffer
14.         {
15.  
16.             get
17.             {
18.                 isUsingA = !isUsingA;
19.                 if (!isUsingA)
20.                 {
21.                     return _swapBufferB;
22.                 }
23.                 else
24.                 {
25.                     return _swapBufferA;
26.                 }
27.             }
28.         }
29.  
30.  
31.         private Settings _settings;
32.  
33.         private static readonly int _TFMID = Shader.PropertyToID("_TFM");
34.  
35.         public KuawaharaRenderPass(Material material)
36.         {
37.             this.profilingSampler = m_ProfilingSampler;
38.             _material = material;
39.             renderPassEvent = RenderPassEvent.BeforeRenderingPostProcessing;
40.         }
41.         public void Setup(in RenderingData renderingData, Settings settings)
42.         {
43.             this._settings = settings;
44.  
45.             ConfigureTarget(renderingData.cameraData.renderer.cameraColorTargetHandle);
46.             ConfigureInput(ScriptableRenderPassInput.Color);
47.             _material.SetInt("_KernelSize", _settings.kernelSize);
48.             _material.SetInt("_N", 8);
49.             _material.SetFloat("_Q", _settings.sharpness);
50.             _material.SetFloat("_Hardness", _settings.hardness);
51.             _material.SetFloat("_Alpha", _settings.alpha);
52.             _material.SetFloat("_ZeroCrossing", _settings.zeroCrossing);
53.             _material.SetFloat("_Zeta", _settings.useZeta ? _settings.zeta : 2.0f / 2.0f / (_settings.kernelSize / 2.0f));
54.         }
55.  
56.  
57.  
58.  
59.         public void Dispose()
60.         {
61.             _swapBufferA?.Release();
62.             _swapBufferB?.Release();
63.             _TFMBuffer?.Release();
64.         }
65.  
66.         public override void OnCameraSetup(CommandBuffer cmd, ref RenderingData renderingData)
67.         {
68.             base.OnCameraSetup(cmd, ref renderingData);
69.             var descriptor = renderingData.cameraData.cameraTargetDescriptor;
70.             descriptor.depthBufferBits = (int)DepthBits.None;
71.             RenderingUtils.ReAllocateIfNeeded(ref _swapBufferA, descriptor, name: "_swapBufferA");
72.             RenderingUtils.ReAllocateIfNeeded(ref _swapBufferB, descriptor, name: "_swapBufferB");
73.             RenderingUtils.ReAllocateIfNeeded(ref _TFMBuffer, descriptor, name: "_tfmBuffer");
74.         }
75.  
76.         public override void Execute(ScriptableRenderContext context, ref RenderingData renderingData)
77.         {
78.             var cameraData = renderingData.cameraData;
79.             var cmd = CommandBufferPool.Get();
80.             // ExecutePass(m_PassData, renderingData.cameraData, renderingData.commandBuffer);
81.             if (_material == null)
82.             {
83.                 Debug.Log("NO material");
84.                 // should not happen as we check it in feature
85.                 return;
86.             }
87.  
88.             if (cameraData.isPreviewCamera)
89.             {
90.                 return;
91.             }
92.  
93.  
94.  
95.             using (new ProfilingScope(cmd, profilingSampler))
96.             {
97.  
98.                 var source = cameraData.renderer.cameraColorTargetHandle;
99.                 Blitter.BlitCameraTexture(cmd, source, _swapBufferA);
100.                 Blitter.BlitCameraTexture(cmd, _swapBufferA, _swapBufferB, _material, 0);
101.                 Blitter.BlitCameraTexture(cmd, _swapBufferB, _swapBufferA, _material, 1);
102.                 Blitter.BlitCameraTexture(cmd, _swapBufferA, _TFMBuffer, _material, 2);
103.  
104.                 _material.SetTexture(_TFMID, _TFMBuffer);
105.  
106.  
107.  
108.                  Blitter.BlitCameraTexture(cmd, source, _swapBufferA, _material, 3);
109.                  Blitter.BlitCameraTexture(cmd, _swapBufferA, _swapBufferB, _material, 3);
110.                  Blitter.BlitCameraTexture(cmd, _swapBufferB, _swapBufferA, _material, 3);
111.                  Blitter.BlitCameraTexture(cmd, _swapBufferA, _swapBufferB, _material, 3);
112.  
113.  
114.                 Blitter.BlitCameraTexture(cmd, _swapBufferB, source);
115.  
116.             }
117.  
118.             context.ExecuteCommandBuffer(cmd);
119.             cmd.Clear();
120.             CommandBufferPool.Release(cmd);
121.  
122.         }
123.  
124.  
125.     }

The thing is: BuiltIn = 150FPS and URP < 20.

I am must be doing something obviously wrong here, the shader code is 99% the same, apart from using URP builtins structures on HLSL side of things. But the performance is absolutely trash.

Am I doing anything wrong?