1// RUN: mlir-opt -test-convert-to-spirv="convert-gpu-modules=true nest-in-gpu-module=true run-signature-conversion=false run-vector-unrolling=false" %s | FileCheck %s 2 3module attributes { 4 gpu.container_module, 5 spirv.target_env = #spirv.target_env<#spirv.vce<v1.0, [Shader], []>, #spirv.resource_limits<>> 6} { 7 // CHECK-LABEL: func.func @main 8 // CHECK: %[[C1:.*]] = arith.constant 1 : index 9 // CHECK: gpu.launch_func @[[$KERNELS_1:.*]]::@[[$BUILTIN_WG_ID_X:.*]] blocks in (%[[C1]], %[[C1]], %[[C1]]) threads in (%[[C1]], %[[C1]], %[[C1]]) 10 func.func @main() { 11 %c1 = arith.constant 1 : index 12 gpu.launch_func @kernels_1::@builtin_workgroup_id_x 13 blocks in (%c1, %c1, %c1) threads in (%c1, %c1, %c1) 14 return 15 } 16 17 // CHECK: gpu.module @[[$KERNELS_1]] 18 // CHECK: spirv.module @{{.*}} Logical GLSL450 19 // CHECK: spirv.func @[[$BUILTIN_WG_ID_X]] 20 // CHECK: spirv.mlir.addressof 21 // CHECK: spirv.Load "Input" 22 // CHECK: spirv.CompositeExtract 23 gpu.module @kernels_1 { 24 gpu.func @builtin_workgroup_id_x() kernel 25 attributes {spirv.entry_point_abi = #spirv.entry_point_abi<workgroup_size = [16, 1, 1]>} { 26 %0 = gpu.block_id x 27 gpu.return 28 } 29 } 30} 31