22 lines · plain
1// RUN: mlir-opt %s \2// RUN: | mlir-opt -gpu-lower-to-nvvm-pipeline="cubin-chip=sm_80 ptxas-cmd-options='-v --register-usage-level=8' allow-pattern-rollback=0" -debug-only=serialize-to-binary \3// RUN: 2>&1 | FileCheck %s4 5func.func @host_function(%arg0 : f32, %arg1 : memref<?xf32>) {6 %cst = arith.constant 1 : index7 %c0 = arith.constant 0 : index8 %cst2 = memref.dim %arg1, %c0 : memref<?xf32>9 10 gpu.launch blocks(%bx, %by, %bz) in (%grid_x = %cst, %grid_y = %cst, %grid_z = %cst)11 threads(%tx, %ty, %tz) in (%block_x = %cst2, %block_y = %cst, %block_z = %cst) {12 memref.store %arg0, %arg1[%tx] : memref<?xf32>13 gpu.terminator14 }15 16 return17}18 19// CHECK: ptxas -arch sm_8020// CHECK-SAME: -v 21// CHECK-SAME: --register-usage-level=822