{
  "domain": "com.microsoft",
  "name": "SkipSimplifiedLayerNormalization",
  "sinceVersion": 1,
  "inputs": {
    "inputT": { "onnx": "input", "dtype": "T" },
    "skipT": { "onnx": "skip", "dtype": "T" },
    "gammaT": { "onnx": "gamma", "dtype": "T", "rank": 1 },
    "biasT": { "onnx": "bias", "dtype": "T", "rank": 1, "optional": true }
  },
  "outputs": {
    "outputT": { "onnx": "output", "dtype": "T", "rank": "ranks.inputT", "shape": "shapes.inputT" },
    "residualT": {
      "onnx": "input_skip_bias_sum",
      "dtype": "T",
      "rank": "ranks.inputT",
      "optional": true,
      "shape": "shapes.inputT"
    }
  },
  "attributes": { "epsilon": { "default": 9.999999960041972e-13 } },
  "typeConstraints": { "T": ["float32", "float16"] },
  "tunables": { "MAX_WORKGROUP_SIZE": { "default": 256 } },
  "derive": {
    "rowCount": "numel(shapes.inputT) / max(1, dim(shapes.inputT, -1))",
    "hiddenSize": "dim(shapes.inputT, -1)",
    "skipWg": "max(1, min(tunables.MAX_WORKGROUP_SIZE, device.limits.maxComputeInvocationsPerWorkgroup, device.limits.maxComputeWorkgroupSizeX, pow2ceil(hiddenSize)))",
    "skipWgVec4": "max(1, min(tunables.MAX_WORKGROUP_SIZE, device.limits.maxComputeInvocationsPerWorkgroup, device.limits.maxComputeWorkgroupSizeX, pow2ceil(ceilDiv(hiddenSize, 4))))",
    "rowDispatchFits": "rowCount <= min(device.limits.maxComputeWorkgroupsPerDimension, 65535) * min(device.limits.maxComputeWorkgroupsPerDimension, 65535)",
    "normResourcesFit": "skipWg * 8 <= device.limits.maxComputeWorkgroupStorageSize and skipWgVec4 * 8 <= device.limits.maxComputeWorkgroupStorageSize",
    "epsilonOk": "attrs.epsilon >= 0",
    "coreContract": "epsilonOk and (ranks.inputT == 2 or ranks.inputT == 3) and ranks.skipT == ranks.inputT and ranks.gammaT == 1 and ranks.outputT == ranks.inputT and sameShape(shapes.inputT, shapes.skipT) and sameShape(shapes.outputT, shapes.inputT) and dim(shapes.inputT, -1) > 0 and dim(shapes.gammaT, 0) == dim(shapes.inputT, -1)",
    "residualOutputContract": "present.residualT and sameShape(shapes.residualT, shapes.inputT)",
    "outputOnlyContract": "not present.residualT",
    "f32MainDtypes": "tensorDtypes.inputT == \"float32\" and tensorDtypes.skipT == \"float32\" and tensorDtypes.gammaT == \"float32\" and tensorDtypes.outputT == \"float32\"",
    "f16MainDtypes": "tensorDtypes.inputT == \"float16\" and tensorDtypes.skipT == \"float16\" and tensorDtypes.gammaT == \"float16\" and tensorDtypes.outputT == \"float16\"",
    "f32ResidualDtypes": "f32MainDtypes and tensorDtypes.residualT == \"float32\" if present.residualT else false",
    "f16ResidualDtypes": "f16MainDtypes and tensorDtypes.residualT == \"float16\" if present.residualT else false",
    "vec4Aligned": "dim(shapes.inputT, -1) % 4 == 0",
    "hasSubgroups": "device.features.has(\"subgroups\") and device.wgslLanguageFeatures.has(\"subgroup_id\")",
    "hasF16": "device.features.has(\"shader-f16\")",
    "no_bias_contract": "not present.biasT",
    "f32_bias_contract": "false if not present.biasT else (ranks.biasT == 1 and tensorDtypes.biasT == \"float32\" and dim(shapes.biasT, 0) == hiddenSize)",
    "f16_bias_contract": "false if not present.biasT else (ranks.biasT == 1 and tensorDtypes.biasT == \"float16\" and dim(shapes.biasT, 0) == hiddenSize)",
    "f32_no_bias_residual_contract": "coreContract and residualOutputContract and f32ResidualDtypes and no_bias_contract",
    "f32_bias_residual_contract": "coreContract and residualOutputContract and f32ResidualDtypes and f32_bias_contract",
    "f16_no_bias_residual_contract": "hasF16 and coreContract and residualOutputContract and f16ResidualDtypes and no_bias_contract",
    "f16_bias_residual_contract": "hasF16 and coreContract and residualOutputContract and f16ResidualDtypes and f16_bias_contract",
    "f32_no_bias_output_contract": "coreContract and outputOnlyContract and f32MainDtypes and no_bias_contract",
    "f32_bias_output_contract": "coreContract and outputOnlyContract and f32MainDtypes and f32_bias_contract",
    "f16_no_bias_output_contract": "hasF16 and coreContract and outputOnlyContract and f16MainDtypes and no_bias_contract",
    "f16_bias_output_contract": "hasF16 and coreContract and outputOnlyContract and f16MainDtypes and f16_bias_contract"
  },
  "when": ["normResourcesFit", "rowDispatchFits"],
  "bindings": {
    "input": { "arg": "inputT", "buffer": "read-only-storage", "elementType": "$vectorScalar" },
    "skip": { "arg": "skipT", "buffer": "read-only-storage", "elementType": "$vectorScalar" },
    "gamma": { "arg": "gammaT", "buffer": "read-only-storage", "elementType": "$vectorScalar", "length": "$HIDDEN_LEN" },
    "output": { "arg": "outputT", "buffer": "storage", "elementType": "$vectorScalar" },
    "input_skip_bias_sum": { "arg": "residualT", "buffer": "storage", "elementType": "$vectorScalar" },
    "params": {
      "buffer": "uniform",
      "struct": [
        { "name": "rows", "type": "u32", "value": "rowCount" },
        {
          "name": "rowStride",
          "type": "u32",
          "value": "max(1, min(rowCount, min(device.limits.maxComputeWorkgroupsPerDimension, 65535)))"
        },
        { "name": "epsilon", "type": "f32", "value": "attrs.epsilon" }
      ]
    },
    "bias": { "arg": "biasT", "buffer": "read-only-storage", "elementType": "$vectorScalar", "length": "$HIDDEN_LEN" },
    "input_2": { "arg": "inputT", "name": "input", "buffer": "read-only-storage", "elementType": "$scalar" },
    "skip_2": { "arg": "skipT", "name": "skip", "buffer": "read-only-storage", "elementType": "$scalar" },
    "gamma_2": {
      "arg": "gammaT",
      "name": "gamma",
      "buffer": "read-only-storage",
      "elementType": "$scalar",
      "length": "$HIDDEN_LEN"
    },
    "output_2": { "arg": "outputT", "name": "output", "buffer": "storage", "elementType": "$scalar" },
    "input_skip_bias_sum_2": {
      "arg": "residualT",
      "name": "input_skip_bias_sum",
      "buffer": "storage",
      "elementType": "$scalar"
    },
    "params_2": {
      "name": "params",
      "buffer": "uniform",
      "struct": [
        { "name": "rows", "type": "u32", "value": "rowCount" },
        { "name": "epsilon", "type": "f32", "value": "attrs.epsilon" }
      ]
    },
    "bias_2": {
      "arg": "biasT",
      "name": "bias",
      "buffer": "read-only-storage",
      "elementType": "$scalar",
      "length": "$HIDDEN_LEN"
    }
  },
  "variants": [
    {
      "id": "no_bias_vec4_f16",
      "priority": 21,
      "when": ["f16_no_bias_residual_contract", "vec4Aligned"],
      "derive": {
        "scalar": "\"f16\"",
        "vectorScalar": "\"vec4<f16>\"",
        "hasBias": "\"no_bias\" == \"bias\"",
        "HIDDEN_LEN": "hiddenSize / 4"
      },
      "passes": [
        {
          "id": "main",
          "name": "SkipSimplifiedLayerNormalization.Vec4",
          "shader": "norm-skip-row-vec4.wgsl.jinja",
          "derive": {
            "simplified": true,
            "hasBias": "\"no_bias\" == \"bias\"",
            "hasBeta": false,
            "writeResidualSum": true,
            "usesF16Spec": true,
            "hidden": "hiddenSize",
            "hiddenVec": "hiddenSize / 4",
            "wg": "skipWgVec4",
            "vecType": "\"vec4<f16>\"",
            "useSubgroups": "hasSubgroups"
          },
          "bindings": ["input", "skip", "gamma", "output", "input_skip_bias_sum", "params"],
          "dispatch": { "x": "min(rowCount, 65535)", "y": "ceilDiv(rowCount, 65535)", "z": 1 },
          "subgroupCollectivesWidth": "portable"
        }
      ]
    },
    {
      "id": "no_bias_vec4",
      "priority": 20,
      "when": ["f32_no_bias_residual_contract", "vec4Aligned"],
      "derive": {
        "scalar": "\"f32\"",
        "vectorScalar": "\"vec4<f32>\"",
        "hasBias": "\"no_bias\" == \"bias\"",
        "HIDDEN_LEN": "hiddenSize / 4"
      },
      "passes": [
        {
          "id": "main",
          "name": "SkipSimplifiedLayerNormalization.Vec4",
          "shader": "norm-skip-row-vec4.wgsl.jinja",
          "derive": {
            "simplified": true,
            "hasBias": "\"no_bias\" == \"bias\"",
            "hasBeta": false,
            "writeResidualSum": true,
            "usesF16Spec": false,
            "hidden": "hiddenSize",
            "hiddenVec": "hiddenSize / 4",
            "wg": "skipWgVec4",
            "vecType": "\"vec4<f32>\"",
            "useSubgroups": "hasSubgroups"
          },
          "bindings": ["input", "skip", "gamma", "output", "input_skip_bias_sum", "params"],
          "dispatch": { "x": "min(rowCount, 65535)", "y": "ceilDiv(rowCount, 65535)", "z": 1 },
          "subgroupCollectivesWidth": "portable"
        }
      ]
    },
    {
      "id": "no_bias_output_only_vec4",
      "priority": 20,
      "when": ["f32_no_bias_output_contract", "vec4Aligned"],
      "derive": {
        "scalar": "\"f32\"",
        "vectorScalar": "\"vec4<f32>\"",
        "hasBias": "\"no_bias\" == \"bias\"",
        "HIDDEN_LEN": "hiddenSize / 4"
      },
      "passes": [
        {
          "id": "main",
          "name": "SkipSimplifiedLayerNormalization.Vec4OutputOnly",
          "shader": "norm-skip-row-vec4.wgsl.jinja",
          "derive": {
            "simplified": true,
            "hasBias": "\"no_bias\" == \"bias\"",
            "hasBeta": false,
            "writeResidualSum": false,
            "usesF16Spec": false,
            "hidden": "hiddenSize",
            "hiddenVec": "hiddenSize / 4",
            "wg": "skipWgVec4",
            "vecType": "\"vec4<f32>\"",
            "useSubgroups": "hasSubgroups"
          },
          "bindings": ["input", "skip", "gamma", "output", "params"],
          "dispatch": { "x": "min(rowCount, 65535)", "y": "ceilDiv(rowCount, 65535)", "z": 1 },
          "subgroupCollectivesWidth": "portable"
        }
      ]
    },
    {
      "id": "no_bias_output_only_vec4_f16",
      "priority": 21,
      "when": ["f16_no_bias_output_contract", "vec4Aligned"],
      "derive": {
        "scalar": "\"f16\"",
        "vectorScalar": "\"vec4<f16>\"",
        "hasBias": "\"no_bias\" == \"bias\"",
        "HIDDEN_LEN": "hiddenSize / 4"
      },
      "passes": [
        {
          "id": "main",
          "name": "SkipSimplifiedLayerNormalization.Vec4OutputOnly",
          "shader": "norm-skip-row-vec4.wgsl.jinja",
          "derive": {
            "simplified": true,
            "hasBias": "\"no_bias\" == \"bias\"",
            "hasBeta": false,
            "writeResidualSum": false,
            "usesF16Spec": true,
            "hidden": "hiddenSize",
            "hiddenVec": "hiddenSize / 4",
            "wg": "skipWgVec4",
            "vecType": "\"vec4<f16>\"",
            "useSubgroups": "hasSubgroups"
          },
          "bindings": ["input", "skip", "gamma", "output", "params"],
          "dispatch": { "x": "min(rowCount, 65535)", "y": "ceilDiv(rowCount, 65535)", "z": 1 },
          "subgroupCollectivesWidth": "portable"
        }
      ]
    },
    {
      "id": "no_bias",
      "priority": 0,
      "when": ["f32_no_bias_residual_contract"],
      "derive": {
        "simplified": true,
        "useSubgroups": false,
        "hasBeta": false,
        "writeResidualSum": true,
        "hasBias": "\"no_bias\" == \"bias\"",
        "scalar": "\"f32\"",
        "workgroupSize": "skipWg",
        "HIDDEN_LEN": "hiddenSize"
      },
      "passes": [
        {
          "id": "main",
          "name": "SkipSimplifiedLayerNormalization",
          "shader": "norm-skip-row.wgsl.jinja",
          "bindings": ["input_2", "skip_2", "gamma_2", "output_2", "input_skip_bias_sum_2", "params_2"],
          "dispatch": { "x": "min(rowCount, 65535)", "y": "ceilDiv(rowCount, 65535)", "z": 1 }
        }
      ]
    },
    {
      "id": "no_bias_f16",
      "priority": 0,
      "when": ["f16_no_bias_residual_contract"],
      "requires": { "features": ["shader-f16"] },
      "derive": {
        "simplified": true,
        "useSubgroups": false,
        "hasBeta": false,
        "writeResidualSum": true,
        "hasBias": "\"no_bias\" == \"bias\"",
        "scalar": "\"f16\"",
        "usesF16": true,
        "workgroupSize": "skipWg",
        "HIDDEN_LEN": "hiddenSize"
      },
      "passes": [
        {
          "id": "main",
          "name": "SkipSimplifiedLayerNormalization",
          "shader": "norm-skip-row.wgsl.jinja",
          "bindings": ["input_2", "skip_2", "gamma_2", "output_2", "input_skip_bias_sum_2", "params_2"],
          "dispatch": { "x": "min(rowCount, 65535)", "y": "ceilDiv(rowCount, 65535)", "z": 1 }
        }
      ]
    },
    {
      "id": "no_bias_output_only_f16",
      "priority": 0,
      "when": ["f16_no_bias_output_contract"],
      "requires": { "features": ["shader-f16"] },
      "derive": {
        "simplified": true,
        "useSubgroups": false,
        "hasBeta": false,
        "writeResidualSum": false,
        "hasBias": "\"no_bias\" == \"bias\"",
        "scalar": "\"f16\"",
        "usesF16": true,
        "workgroupSize": "skipWg",
        "HIDDEN_LEN": "hiddenSize"
      },
      "passes": [
        {
          "id": "main",
          "name": "SkipSimplifiedLayerNormalization.OutputOnly",
          "shader": "norm-skip-row.wgsl.jinja",
          "bindings": ["input_2", "skip_2", "gamma_2", "output_2", "params_2"],
          "dispatch": { "x": "min(rowCount, 65535)", "y": "ceilDiv(rowCount, 65535)", "z": 1 }
        }
      ]
    },
    {
      "id": "no_bias_output_only",
      "priority": 0,
      "when": ["f32_no_bias_output_contract"],
      "derive": {
        "simplified": true,
        "useSubgroups": false,
        "hasBeta": false,
        "writeResidualSum": false,
        "hasBias": "\"no_bias\" == \"bias\"",
        "scalar": "\"f32\"",
        "workgroupSize": "skipWg",
        "HIDDEN_LEN": "hiddenSize"
      },
      "passes": [
        {
          "id": "main",
          "name": "SkipSimplifiedLayerNormalization.OutputOnly",
          "shader": "norm-skip-row.wgsl.jinja",
          "bindings": ["input_2", "skip_2", "gamma_2", "output_2", "params_2"],
          "dispatch": { "x": "min(rowCount, 65535)", "y": "ceilDiv(rowCount, 65535)", "z": 1 }
        }
      ]
    },
    {
      "id": "bias_vec4_f16",
      "priority": 21,
      "when": ["f16_bias_residual_contract", "vec4Aligned"],
      "derive": {
        "scalar": "\"f16\"",
        "vectorScalar": "\"vec4<f16>\"",
        "hasBias": "\"bias\" == \"bias\"",
        "HIDDEN_LEN": "hiddenSize / 4"
      },
      "passes": [
        {
          "id": "main",
          "name": "SkipSimplifiedLayerNormalization.Vec4",
          "shader": "norm-skip-row-vec4.wgsl.jinja",
          "derive": {
            "simplified": true,
            "hasBias": "\"bias\" == \"bias\"",
            "hasBeta": false,
            "writeResidualSum": true,
            "usesF16Spec": true,
            "hidden": "hiddenSize",
            "hiddenVec": "hiddenSize / 4",
            "wg": "skipWgVec4",
            "vecType": "\"vec4<f16>\"",
            "useSubgroups": "hasSubgroups"
          },
          "bindings": ["input", "skip", "gamma", "bias", "output", "input_skip_bias_sum", "params"],
          "dispatch": { "x": "min(rowCount, 65535)", "y": "ceilDiv(rowCount, 65535)", "z": 1 },
          "subgroupCollectivesWidth": "portable"
        }
      ]
    },
    {
      "id": "bias_vec4",
      "priority": 20,
      "when": ["f32_bias_residual_contract", "vec4Aligned"],
      "derive": {
        "scalar": "\"f32\"",
        "vectorScalar": "\"vec4<f32>\"",
        "hasBias": "\"bias\" == \"bias\"",
        "HIDDEN_LEN": "hiddenSize / 4"
      },
      "passes": [
        {
          "id": "main",
          "name": "SkipSimplifiedLayerNormalization.Vec4",
          "shader": "norm-skip-row-vec4.wgsl.jinja",
          "derive": {
            "simplified": true,
            "hasBias": "\"bias\" == \"bias\"",
            "hasBeta": false,
            "writeResidualSum": true,
            "usesF16Spec": false,
            "hidden": "hiddenSize",
            "hiddenVec": "hiddenSize / 4",
            "wg": "skipWgVec4",
            "vecType": "\"vec4<f32>\"",
            "useSubgroups": "hasSubgroups"
          },
          "bindings": ["input", "skip", "gamma", "bias", "output", "input_skip_bias_sum", "params"],
          "dispatch": { "x": "min(rowCount, 65535)", "y": "ceilDiv(rowCount, 65535)", "z": 1 },
          "subgroupCollectivesWidth": "portable"
        }
      ]
    },
    {
      "id": "bias_output_only_vec4",
      "priority": 20,
      "when": ["f32_bias_output_contract", "vec4Aligned"],
      "derive": {
        "scalar": "\"f32\"",
        "vectorScalar": "\"vec4<f32>\"",
        "hasBias": "\"bias\" == \"bias\"",
        "HIDDEN_LEN": "hiddenSize / 4"
      },
      "passes": [
        {
          "id": "main",
          "name": "SkipSimplifiedLayerNormalization.Vec4OutputOnly",
          "shader": "norm-skip-row-vec4.wgsl.jinja",
          "derive": {
            "simplified": true,
            "hasBias": "\"bias\" == \"bias\"",
            "hasBeta": false,
            "writeResidualSum": false,
            "usesF16Spec": false,
            "hidden": "hiddenSize",
            "hiddenVec": "hiddenSize / 4",
            "wg": "skipWgVec4",
            "vecType": "\"vec4<f32>\"",
            "useSubgroups": "hasSubgroups"
          },
          "bindings": ["input", "skip", "gamma", "bias", "output", "params"],
          "dispatch": { "x": "min(rowCount, 65535)", "y": "ceilDiv(rowCount, 65535)", "z": 1 },
          "subgroupCollectivesWidth": "portable"
        }
      ]
    },
    {
      "id": "bias_output_only_vec4_f16",
      "priority": 21,
      "when": ["f16_bias_output_contract", "vec4Aligned"],
      "derive": {
        "scalar": "\"f16\"",
        "vectorScalar": "\"vec4<f16>\"",
        "hasBias": "\"bias\" == \"bias\"",
        "HIDDEN_LEN": "hiddenSize / 4"
      },
      "passes": [
        {
          "id": "main",
          "name": "SkipSimplifiedLayerNormalization.Vec4OutputOnly",
          "shader": "norm-skip-row-vec4.wgsl.jinja",
          "derive": {
            "simplified": true,
            "hasBias": "\"bias\" == \"bias\"",
            "hasBeta": false,
            "writeResidualSum": false,
            "usesF16Spec": true,
            "hidden": "hiddenSize",
            "hiddenVec": "hiddenSize / 4",
            "wg": "skipWgVec4",
            "vecType": "\"vec4<f16>\"",
            "useSubgroups": "hasSubgroups"
          },
          "bindings": ["input", "skip", "gamma", "bias", "output", "params"],
          "dispatch": { "x": "min(rowCount, 65535)", "y": "ceilDiv(rowCount, 65535)", "z": 1 },
          "subgroupCollectivesWidth": "portable"
        }
      ]
    },
    {
      "id": "bias",
      "priority": 0,
      "when": ["f32_bias_residual_contract"],
      "derive": {
        "simplified": true,
        "useSubgroups": false,
        "hasBeta": false,
        "writeResidualSum": true,
        "hasBias": "\"bias\" == \"bias\"",
        "scalar": "\"f32\"",
        "workgroupSize": "skipWg",
        "HIDDEN_LEN": "hiddenSize"
      },
      "passes": [
        {
          "id": "main",
          "name": "SkipSimplifiedLayerNormalization",
          "shader": "norm-skip-row.wgsl.jinja",
          "bindings": ["input_2", "skip_2", "gamma_2", "bias_2", "output_2", "input_skip_bias_sum_2", "params_2"],
          "dispatch": { "x": "min(rowCount, 65535)", "y": "ceilDiv(rowCount, 65535)", "z": 1 }
        }
      ]
    },
    {
      "id": "bias_f16",
      "priority": 0,
      "when": ["f16_bias_residual_contract"],
      "requires": { "features": ["shader-f16"] },
      "derive": {
        "simplified": true,
        "useSubgroups": false,
        "hasBeta": false,
        "writeResidualSum": true,
        "hasBias": "\"bias\" == \"bias\"",
        "scalar": "\"f16\"",
        "usesF16": true,
        "workgroupSize": "skipWg",
        "HIDDEN_LEN": "hiddenSize"
      },
      "passes": [
        {
          "id": "main",
          "name": "SkipSimplifiedLayerNormalization",
          "shader": "norm-skip-row.wgsl.jinja",
          "bindings": ["input_2", "skip_2", "gamma_2", "bias_2", "output_2", "input_skip_bias_sum_2", "params_2"],
          "dispatch": { "x": "min(rowCount, 65535)", "y": "ceilDiv(rowCount, 65535)", "z": 1 }
        }
      ]
    },
    {
      "id": "bias_output_only_f16",
      "priority": 0,
      "when": ["f16_bias_output_contract"],
      "requires": { "features": ["shader-f16"] },
      "derive": {
        "simplified": true,
        "useSubgroups": false,
        "hasBeta": false,
        "writeResidualSum": false,
        "hasBias": "\"bias\" == \"bias\"",
        "scalar": "\"f16\"",
        "usesF16": true,
        "workgroupSize": "skipWg",
        "HIDDEN_LEN": "hiddenSize"
      },
      "passes": [
        {
          "id": "main",
          "name": "SkipSimplifiedLayerNormalization.OutputOnly",
          "shader": "norm-skip-row.wgsl.jinja",
          "bindings": ["input_2", "skip_2", "gamma_2", "bias_2", "output_2", "params_2"],
          "dispatch": { "x": "min(rowCount, 65535)", "y": "ceilDiv(rowCount, 65535)", "z": 1 }
        }
      ]
    },
    {
      "id": "bias_output_only",
      "priority": 0,
      "when": ["f32_bias_output_contract"],
      "derive": {
        "simplified": true,
        "useSubgroups": false,
        "hasBeta": false,
        "writeResidualSum": false,
        "hasBias": "\"bias\" == \"bias\"",
        "scalar": "\"f32\"",
        "workgroupSize": "skipWg",
        "HIDDEN_LEN": "hiddenSize"
      },
      "passes": [
        {
          "id": "main",
          "name": "SkipSimplifiedLayerNormalization.OutputOnly",
          "shader": "norm-skip-row.wgsl.jinja",
          "bindings": ["input_2", "skip_2", "gamma_2", "bias_2", "output_2", "params_2"],
          "dispatch": { "x": "min(rowCount, 65535)", "y": "ceilDiv(rowCount, 65535)", "z": 1 }
        }
      ]
    }
  ]
}