brintos

brintos / llvm-project-archived public Read only

0
0
Text · 2.3 KiB · b6ef008 Raw
53 lines · plain
1// RUN: mlir-opt -allow-unregistered-dialect -pass-pipeline="builtin.module(func.func(test-mapping-to-processing-elements))" %s | FileCheck %s2 3// CHECK: #[[mul_map:.+]] = affine_map<()[s0, s1] -> (s0 * s1)>4// CHECK: #[[add_map:.+]] = affine_map<()[s0, s1] -> (s0 + s1)>5 6//      CHECK: func @map1d7// CHECK-SAME: (%[[lb:.*]]: index, %[[ub:.*]]: index, %[[step:.*]]: index)8func.func @map1d(%lb: index, %ub: index, %step: index) {9  // CHECK: %[[threads:.*]]:2 = "new_processor_id_and_range"() : () -> (index, index)10  %0:2 = "new_processor_id_and_range"() : () -> (index, index)11 12  // CHECK: %[[thread_offset:.+]] = affine.apply #[[mul_map]]()[%[[threads]]#0, %[[step]]]13  // CHECK: %[[new_lb:.+]] = affine.apply #[[add_map]]()[%[[thread_offset]], %[[lb]]]14  // CHECK: %[[new_step:.+]] = affine.apply #[[mul_map]]()[%[[threads]]#1, %[[step]]]15 16  // CHECK: scf.for %{{.*}} = %[[new_lb]] to %[[ub]] step %[[new_step]] {17  scf.for %i = %lb to %ub step %step {}18  return19}20 21//      CHECK: func @map2d22// CHECK-SAME: (%[[lb:.*]]: index, %[[ub:.*]]: index, %[[step:.*]]: index)23func.func @map2d(%lb : index, %ub : index, %step : index) {24  // CHECK: %[[blocks:.*]]:2 = "new_processor_id_and_range"() : () -> (index, index)25  %0:2 = "new_processor_id_and_range"() : () -> (index, index)26 27  // CHECK: %[[threads:.*]]:2 = "new_processor_id_and_range"() : () -> (index, index)28  %1:2 = "new_processor_id_and_range"() : () -> (index, index)29 30  // blockIdx.x * blockDim.x31  // CHECK: %[[bidxXbdimx:.+]] = affine.apply #[[mul_map]]()[%[[blocks]]#0, %[[threads]]#1]32  //33  // threadIdx.x + blockIdx.x * blockDim.x34  // CHECK: %[[tidxpbidxXbdimx:.+]] = affine.apply #[[add_map]]()[%[[bidxXbdimx]], %[[threads]]#0]35  //36  // thread_offset = step * (threadIdx.x + blockIdx.x * blockDim.x)37  // CHECK: %[[thread_offset:.+]] = affine.apply #[[mul_map]]()[%[[tidxpbidxXbdimx]], %[[step]]]38  //39  // new_lb = lb + thread_offset40  // CHECK: %[[new_lb:.+]] = affine.apply #[[add_map]]()[%[[thread_offset]], %[[lb]]]41  //42  // stepXgdimx = step * gridDim.x43  // CHECK: %[[stepXgdimx:.+]] = affine.apply #[[mul_map]]()[%[[blocks]]#1, %[[step]]]44  //45  // new_step = step * gridDim.x * blockDim.x46  // CHECK: %[[new_step:.+]] = affine.apply #[[mul_map]]()[%[[threads]]#1, %[[stepXgdimx]]]47  //48  // CHECK: scf.for %{{.*}} = %[[new_lb]] to %[[ub]] step %[[new_step]] {49 50  scf.for %i = %lb to %ub step %step {}51  return52}53