Google Git
Sign in
llvm / llvm-project / c41b318423c4dbf0a65f81e5e9a816c1710ba4f6 / . / mlir / test / Dialect / SparseTensor / sparse_1d.mlir
blob: 1990803634f5df5c3dbec56b9bdd405261afdbc5 [file] [log] [blame]
// NOTE: Assertions have been autogenerated by utils/generate-test-checks.py
// RUN: mlir-opt %s -sparsification | FileCheck %s
#DV = #sparse_tensor.encoding<{ dimLevelType = [ "dense" ] }>
#SV = #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ] }>
#trait1 = {
indexing_maps = [
affine_map<(i) -> (i)>, // a
affine_map<(i) -> (i)> // x (out)
],
iterator_types = ["parallel"],
doc = "x(i) = a(i) OP b"
}
// CHECK-LABEL: func @add_d(
// CHECK-SAME: %[[VAL_0:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_1:.*]]: f32,
// CHECK-SAME: %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
// CHECK: %[[VAL_3:.*]] = arith.constant 32 : index
// CHECK: %[[VAL_4:.*]] = arith.constant 0 : index
// CHECK: %[[VAL_5:.*]] = arith.constant 1 : index
// CHECK: %[[VAL_6:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_7:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
// CHECK: %[[VAL_8:.*]] = memref.alloc() : memref<32xf32>
// CHECK: memref.copy %[[VAL_7]], %[[VAL_8]] : memref<32xf32> to memref<32xf32>
// CHECK: scf.for %[[VAL_9:.*]] = %[[VAL_4]] to %[[VAL_3]] step %[[VAL_5]] {
// CHECK: %[[VAL_10:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_9]]] : memref<?xf32>
// CHECK: %[[VAL_11:.*]] = arith.addf %[[VAL_10]], %[[VAL_1]] : f32
// CHECK: memref.store %[[VAL_11]], %[[VAL_8]]{{\[}}%[[VAL_9]]] : memref<32xf32>
// CHECK: }
// CHECK: %[[VAL_12:.*]] = bufferization.to_tensor %[[VAL_8]] : memref<32xf32>
// CHECK: return %[[VAL_12]] : tensor<32xf32>
// CHECK: }
func @add_d(%arga: tensor<32xf32, #DV>, %argb: f32, %argx: tensor<32xf32>) -> tensor<32xf32> {
%0 = linalg.generic #trait1
ins(%arga: tensor<32xf32, #DV>)
outs(%argx: tensor<32xf32>) {
^bb(%a: f32, %x: f32):
%0 = arith.addf %a, %argb : f32
linalg.yield %0 : f32
} -> tensor<32xf32>
return %0 : tensor<32xf32>
}
// CHECK-LABEL: func @add_d_init(
// CHECK-SAME: %[[VAL_0:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_1:.*]]: f32) -> tensor<32xf32> {
// CHECK: %[[VAL_2:.*]] = arith.constant 32 : index
// CHECK: %[[VAL_3:.*]] = arith.constant 0.000000e+00 : f32
// CHECK: %[[VAL_4:.*]] = arith.constant 0 : index
// CHECK: %[[VAL_5:.*]] = arith.constant 1 : index
// CHECK: %[[VAL_6:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_7:.*]] = memref.alloc() : memref<32xf32>
// CHECK: linalg.fill(%[[VAL_3]], %[[VAL_7]]) : f32, memref<32xf32>
// CHECK: scf.for %[[VAL_8:.*]] = %[[VAL_4]] to %[[VAL_2]] step %[[VAL_5]] {
// CHECK: %[[VAL_9:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_8]]] : memref<?xf32>
// CHECK: %[[VAL_10:.*]] = arith.addf %[[VAL_9]], %[[VAL_1]] : f32
// CHECK: memref.store %[[VAL_10]], %[[VAL_7]]{{\[}}%[[VAL_8]]] : memref<32xf32>
// CHECK: }
// CHECK: %[[VAL_11:.*]] = bufferization.to_tensor %[[VAL_7]] : memref<32xf32>
// CHECK: return %[[VAL_11]] : tensor<32xf32>
// CHECK: }
func @add_d_init(%arga: tensor<32xf32, #DV>, %argb: f32) -> tensor<32xf32> {
%u = linalg.init_tensor [32] : tensor<32xf32>
%0 = linalg.generic #trait1
ins(%arga: tensor<32xf32, #DV>)
outs(%u: tensor<32xf32>) {
^bb(%a: f32, %x: f32):
%0 = arith.addf %a, %argb : f32
linalg.yield %0 : f32
} -> tensor<32xf32>
return %0 : tensor<32xf32>
}
// CHECK-LABEL: func @mul_d(
// CHECK-SAME: %[[VAL_0:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_1:.*]]: f32,
// CHECK-SAME: %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
// CHECK: %[[VAL_3:.*]] = arith.constant 32 : index
// CHECK: %[[VAL_4:.*]] = arith.constant 0 : index
// CHECK: %[[VAL_5:.*]] = arith.constant 1 : index
// CHECK: %[[VAL_6:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_7:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
// CHECK: %[[VAL_8:.*]] = memref.alloc() : memref<32xf32>
// CHECK: memref.copy %[[VAL_7]], %[[VAL_8]] : memref<32xf32> to memref<32xf32>
// CHECK: scf.for %[[VAL_9:.*]] = %[[VAL_4]] to %[[VAL_3]] step %[[VAL_5]] {
// CHECK: %[[VAL_10:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_9]]] : memref<?xf32>
// CHECK: %[[VAL_11:.*]] = arith.mulf %[[VAL_10]], %[[VAL_1]] : f32
// CHECK: memref.store %[[VAL_11]], %[[VAL_8]]{{\[}}%[[VAL_9]]] : memref<32xf32>
// CHECK: }
// CHECK: %[[VAL_12:.*]] = bufferization.to_tensor %[[VAL_8]] : memref<32xf32>
// CHECK: return %[[VAL_12]] : tensor<32xf32>
// CHECK: }
func @mul_d(%arga: tensor<32xf32, #DV>, %argb: f32, %argx: tensor<32xf32>) -> tensor<32xf32> {
%0 = linalg.generic #trait1
ins(%arga: tensor<32xf32, #DV>)
outs(%argx: tensor<32xf32>) {
^bb(%a: f32, %x: f32):
%0 = arith.mulf %a, %argb : f32
linalg.yield %0 : f32
} -> tensor<32xf32>
return %0 : tensor<32xf32>
}
// CHECK-LABEL: func @add_s(
// CHECK-SAME: %[[VAL_0:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_1:.*]]: f32,
// CHECK-SAME: %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
// CHECK-DAG: %[[VAL_3:.*]] = arith.constant 32 : index
// CHECK-DAG: %[[VAL_4:.*]] = arith.constant 0 : index
// CHECK-DAG: %[[VAL_5:.*]] = arith.constant true
// CHECK-DAG: %[[VAL_6:.*]] = arith.constant 1 : index
// CHECK: %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
// CHECK: %[[VAL_11:.*]] = memref.alloc() : memref<32xf32>
// CHECK: memref.copy %[[VAL_10]], %[[VAL_11]] : memref<32xf32> to memref<32xf32>
// CHECK: %[[VAL_12:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_4]]] : memref<?xindex>
// CHECK: %[[VAL_13:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_6]]] : memref<?xindex>
// CHECK: %[[VAL_14:.*]]:2 = scf.while (%[[VAL_15:.*]] = %[[VAL_12]], %[[VAL_16:.*]] = %[[VAL_4]]) : (index, index) -> (index, index) {
// CHECK: %[[VAL_17:.*]] = arith.cmpi ult, %[[VAL_15]], %[[VAL_13]] : index
// CHECK: scf.condition(%[[VAL_17]]) %[[VAL_15]], %[[VAL_16]] : index, index
// CHECK: } do {
// CHECK: ^bb0(%[[VAL_18:.*]]: index, %[[VAL_19:.*]]: index):
// CHECK: %[[VAL_20:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_18]]] : memref<?xindex>
// CHECK: %[[VAL_21:.*]] = arith.cmpi eq, %[[VAL_20]], %[[VAL_19]] : index
// CHECK: scf.if %[[VAL_21]] {
// CHECK: %[[VAL_22:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_18]]] : memref<?xf32>
// CHECK: %[[VAL_23:.*]] = arith.addf %[[VAL_22]], %[[VAL_1]] : f32
// CHECK: memref.store %[[VAL_23]], %[[VAL_11]]{{\[}}%[[VAL_19]]] : memref<32xf32>
// CHECK: } else {
// CHECK: scf.if %[[VAL_5]] {
// CHECK: memref.store %[[VAL_1]], %[[VAL_11]]{{\[}}%[[VAL_19]]] : memref<32xf32>
// CHECK: } else {
// CHECK: }
// CHECK: }
// CHECK: %[[VAL_24:.*]] = arith.cmpi eq, %[[VAL_20]], %[[VAL_19]] : index
// CHECK: %[[VAL_25:.*]] = arith.addi %[[VAL_18]], %[[VAL_6]] : index
// CHECK: %[[VAL_26:.*]] = select %[[VAL_24]], %[[VAL_25]], %[[VAL_18]] : index
// CHECK: %[[VAL_27:.*]] = arith.addi %[[VAL_19]], %[[VAL_6]] : index
// CHECK: scf.yield %[[VAL_26]], %[[VAL_27]] : index, index
// CHECK: }
// CHECK: scf.for %[[VAL_28:.*]] = %[[VAL_29:.*]]#1 to %[[VAL_3]] step %[[VAL_6]] {
// CHECK: memref.store %[[VAL_1]], %[[VAL_11]]{{\[}}%[[VAL_28]]] : memref<32xf32>
// CHECK: }
// CHECK: %[[VAL_30:.*]] = bufferization.to_tensor %[[VAL_11]] : memref<32xf32>
// CHECK: return %[[VAL_30]] : tensor<32xf32>
// CHECK: }
func @add_s(%arga: tensor<32xf32, #SV>, %argb: f32, %argx: tensor<32xf32>) -> tensor<32xf32> {
%0 = linalg.generic #trait1
ins(%arga: tensor<32xf32, #SV>)
outs(%argx: tensor<32xf32>) {
^bb(%a: f32, %x: f32):
%0 = arith.addf %a, %argb : f32
linalg.yield %0 : f32
} -> tensor<32xf32>
return %0 : tensor<32xf32>
}
// CHECK-LABEL: func @repeated_add_s(
// CHECK-SAME: %[[VAL_0:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_1:.*]]: tensor<32xf32>) -> tensor<32xf32> {
// CHECK: %[[VAL_2:.*]] = arith.constant 0 : index
// CHECK: %[[VAL_3:.*]] = arith.constant 1 : index
// CHECK: %[[VAL_4:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_2]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_5:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_2]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_6:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_7:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xf32>
// CHECK: %[[VAL_8:.*]] = memref.alloc() : memref<32xf32>
// CHECK: memref.copy %[[VAL_7]], %[[VAL_8]] : memref<32xf32> to memref<32xf32>
// CHECK: %[[VAL_9:.*]] = memref.load %[[VAL_4]]{{\[}}%[[VAL_2]]] : memref<?xindex>
// CHECK: %[[VAL_10:.*]] = memref.load %[[VAL_4]]{{\[}}%[[VAL_3]]] : memref<?xindex>
// CHECK: scf.for %[[VAL_11:.*]] = %[[VAL_9]] to %[[VAL_10]] step %[[VAL_3]] {
// CHECK: %[[VAL_12:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_11]]] : memref<?xindex>
// CHECK: %[[VAL_13:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_11]]] : memref<?xf32>
// CHECK: %[[VAL_14:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_11]]] : memref<?xf32>
// CHECK: %[[VAL_15:.*]] = arith.addf %[[VAL_13]], %[[VAL_14]] : f32
// CHECK: %[[VAL_16:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_11]]] : memref<?xf32>
// CHECK: %[[VAL_17:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_11]]] : memref<?xf32>
// CHECK: %[[VAL_18:.*]] = arith.addf %[[VAL_16]], %[[VAL_17]] : f32
// CHECK: %[[VAL_19:.*]] = arith.addf %[[VAL_15]], %[[VAL_18]] : f32
// CHECK: memref.store %[[VAL_19]], %[[VAL_8]]{{\[}}%[[VAL_12]]] : memref<32xf32>
// CHECK: }
// CHECK: %[[VAL_20:.*]] = bufferization.to_tensor %[[VAL_8]] : memref<32xf32>
// CHECK: return %[[VAL_20]] : tensor<32xf32>
// CHECK: }
func @repeated_add_s(%arga: tensor<32xf32, #SV>, %argx: tensor<32xf32>) -> tensor<32xf32> {
%0 = linalg.generic #trait1
ins(%arga: tensor<32xf32, #SV>)
outs(%argx: tensor<32xf32>) {
^bb(%a: f32, %x: f32):
%0 = arith.addf %a, %a : f32 // same tensor
%1 = arith.addf %a, %a : f32 // should yield
%2 = arith.addf %0, %1 : f32 // one guard
linalg.yield %2 : f32
} -> tensor<32xf32>
return %0 : tensor<32xf32>
}
// CHECK-LABEL: func @mul_s(
// CHECK-SAME: %[[VAL_0:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_1:.*]]: f32,
// CHECK-SAME: %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
// CHECK: %[[VAL_3:.*]] = arith.constant 0 : index
// CHECK: %[[VAL_4:.*]] = arith.constant 1 : index
// CHECK: %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
// CHECK: %[[VAL_9:.*]] = memref.alloc() : memref<32xf32>
// CHECK: memref.copy %[[VAL_8]], %[[VAL_9]] : memref<32xf32> to memref<32xf32>
// CHECK: %[[VAL_10:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
// CHECK: %[[VAL_11:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
// CHECK: scf.for %[[VAL_12:.*]] = %[[VAL_10]] to %[[VAL_11]] step %[[VAL_4]] {
// CHECK: %[[VAL_13:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_12]]] : memref<?xindex>
// CHECK: %[[VAL_14:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_12]]] : memref<?xf32>
// CHECK: %[[VAL_15:.*]] = arith.mulf %[[VAL_14]], %[[VAL_1]] : f32
// CHECK: memref.store %[[VAL_15]], %[[VAL_9]]{{\[}}%[[VAL_13]]] : memref<32xf32>
// CHECK: }
// CHECK: %[[VAL_16:.*]] = bufferization.to_tensor %[[VAL_9]] : memref<32xf32>
// CHECK: return %[[VAL_16]] : tensor<32xf32>
// CHECK: }
func @mul_s(%arga: tensor<32xf32, #SV>, %argb: f32, %argx: tensor<32xf32>) -> tensor<32xf32> {
%0 = linalg.generic #trait1
ins(%arga: tensor<32xf32, #SV>)
outs(%argx: tensor<32xf32>) {
^bb(%a: f32, %x: f32):
%0 = arith.mulf %a, %argb : f32
linalg.yield %0 : f32
} -> tensor<32xf32>
return %0 : tensor<32xf32>
}
#trait2 = {
indexing_maps = [
affine_map<(i) -> (i)>, // a
affine_map<(i) -> (i)>, // b
affine_map<(i) -> (i)> // x (out)
],
iterator_types = ["parallel"],
doc = "x(i) = a(i) OP b(i)"
}
// CHECK-LABEL: func @add_dd(
// CHECK-SAME: %[[VAL_0:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_1:.*]]: tensor<32xf32>,
// CHECK-SAME: %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
// CHECK: %[[VAL_3:.*]] = arith.constant 32 : index
// CHECK: %[[VAL_4:.*]] = arith.constant 0 : index
// CHECK: %[[VAL_5:.*]] = arith.constant 1 : index
// CHECK: %[[VAL_6:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_7:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xf32>
// CHECK: %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
// CHECK: %[[VAL_9:.*]] = memref.alloc() : memref<32xf32>
// CHECK: memref.copy %[[VAL_8]], %[[VAL_9]] : memref<32xf32> to memref<32xf32>
// CHECK: scf.for %[[VAL_10:.*]] = %[[VAL_4]] to %[[VAL_3]] step %[[VAL_5]] {
// CHECK: %[[VAL_11:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_10]]] : memref<?xf32>
// CHECK: %[[VAL_12:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_10]]] : memref<32xf32>
// CHECK: %[[VAL_13:.*]] = arith.addf %[[VAL_11]], %[[VAL_12]] : f32
// CHECK: memref.store %[[VAL_13]], %[[VAL_9]]{{\[}}%[[VAL_10]]] : memref<32xf32>
// CHECK: }
// CHECK: %[[VAL_14:.*]] = bufferization.to_tensor %[[VAL_9]] : memref<32xf32>
// CHECK: return %[[VAL_14]] : tensor<32xf32>
// CHECK: }
func @add_dd(%arga: tensor<32xf32, #DV>, %argb: tensor<32xf32>, %argx: tensor<32xf32>) -> tensor<32xf32> {
%0 = linalg.generic #trait2
ins(%arga, %argb: tensor<32xf32, #DV>, tensor<32xf32>)
outs(%argx: tensor<32xf32>) {
^bb(%a: f32, %b: f32, %x: f32):
%0 = arith.addf %a, %b : f32
linalg.yield %0 : f32
} -> tensor<32xf32>
return %0 : tensor<32xf32>
}
// CHECK-LABEL: func @mul_dd(
// CHECK-SAME: %[[VAL_0:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_1:.*]]: tensor<32xf32>,
// CHECK-SAME: %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
// CHECK: %[[VAL_3:.*]] = arith.constant 32 : index
// CHECK: %[[VAL_4:.*]] = arith.constant 0 : index
// CHECK: %[[VAL_5:.*]] = arith.constant 1 : index
// CHECK: %[[VAL_6:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "dense" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_7:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xf32>
// CHECK: %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
// CHECK: %[[VAL_9:.*]] = memref.alloc() : memref<32xf32>
// CHECK: memref.copy %[[VAL_8]], %[[VAL_9]] : memref<32xf32> to memref<32xf32>
// CHECK: scf.for %[[VAL_10:.*]] = %[[VAL_4]] to %[[VAL_3]] step %[[VAL_5]] {
// CHECK: %[[VAL_11:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_10]]] : memref<?xf32>
// CHECK: %[[VAL_12:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_10]]] : memref<32xf32>
// CHECK: %[[VAL_13:.*]] = arith.mulf %[[VAL_11]], %[[VAL_12]] : f32
// CHECK: memref.store %[[VAL_13]], %[[VAL_9]]{{\[}}%[[VAL_10]]] : memref<32xf32>
// CHECK: }
// CHECK: %[[VAL_14:.*]] = bufferization.to_tensor %[[VAL_9]] : memref<32xf32>
// CHECK: return %[[VAL_14]] : tensor<32xf32>
// CHECK: }
func @mul_dd(%arga: tensor<32xf32, #DV>, %argb: tensor<32xf32>, %argx: tensor<32xf32>) -> tensor<32xf32> {
%0 = linalg.generic #trait2
ins(%arga, %argb: tensor<32xf32, #DV>, tensor<32xf32>)
outs(%argx: tensor<32xf32>) {
^bb(%a: f32, %b: f32, %x: f32):
%0 = arith.mulf %a, %b : f32
linalg.yield %0 : f32
} -> tensor<32xf32>
return %0 : tensor<32xf32>
}
// CHECK-LABEL: func @add_ds(
// CHECK-SAME: %[[VAL_0:.*]]: tensor<32xf32>,
// CHECK-SAME: %[[VAL_1:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
// CHECK-DAG: %[[VAL_3:.*]] = arith.constant 32 : index
// CHECK-DAG: %[[VAL_4:.*]] = arith.constant 0 : index
// CHECK-DAG: %[[VAL_5:.*]] = arith.constant true
// CHECK-DAG: %[[VAL_6:.*]] = arith.constant 1 : index
// CHECK: %[[VAL_7:.*]] = bufferization.to_memref %[[VAL_0]] : memref<32xf32>
// CHECK: %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
// CHECK: %[[VAL_12:.*]] = memref.alloc() : memref<32xf32>
// CHECK: memref.copy %[[VAL_11]], %[[VAL_12]] : memref<32xf32> to memref<32xf32>
// CHECK: %[[VAL_13:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_4]]] : memref<?xindex>
// CHECK: %[[VAL_14:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_6]]] : memref<?xindex>
// CHECK: %[[VAL_15:.*]]:2 = scf.while (%[[VAL_16:.*]] = %[[VAL_13]], %[[VAL_17:.*]] = %[[VAL_4]]) : (index, index) -> (index, index) {
// CHECK: %[[VAL_18:.*]] = arith.cmpi ult, %[[VAL_16]], %[[VAL_14]] : index
// CHECK: scf.condition(%[[VAL_18]]) %[[VAL_16]], %[[VAL_17]] : index, index
// CHECK: } do {
// CHECK: ^bb0(%[[VAL_19:.*]]: index, %[[VAL_20:.*]]: index):
// CHECK: %[[VAL_21:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_19]]] : memref<?xindex>
// CHECK: %[[VAL_22:.*]] = arith.cmpi eq, %[[VAL_21]], %[[VAL_20]] : index
// CHECK: scf.if %[[VAL_22]] {
// CHECK: %[[VAL_23:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_20]]] : memref<32xf32>
// CHECK: %[[VAL_24:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_19]]] : memref<?xf32>
// CHECK: %[[VAL_25:.*]] = arith.addf %[[VAL_23]], %[[VAL_24]] : f32
// CHECK: memref.store %[[VAL_25]], %[[VAL_12]]{{\[}}%[[VAL_20]]] : memref<32xf32>
// CHECK: } else {
// CHECK: scf.if %[[VAL_5]] {
// CHECK: %[[VAL_26:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_20]]] : memref<32xf32>
// CHECK: memref.store %[[VAL_26]], %[[VAL_12]]{{\[}}%[[VAL_20]]] : memref<32xf32>
// CHECK: } else {
// CHECK: }
// CHECK: }
// CHECK: %[[VAL_27:.*]] = arith.cmpi eq, %[[VAL_21]], %[[VAL_20]] : index
// CHECK: %[[VAL_28:.*]] = arith.addi %[[VAL_19]], %[[VAL_6]] : index
// CHECK: %[[VAL_29:.*]] = select %[[VAL_27]], %[[VAL_28]], %[[VAL_19]] : index
// CHECK: %[[VAL_30:.*]] = arith.addi %[[VAL_20]], %[[VAL_6]] : index
// CHECK: scf.yield %[[VAL_29]], %[[VAL_30]] : index, index
// CHECK: }
// CHECK: scf.for %[[VAL_31:.*]] = %[[VAL_32:.*]]#1 to %[[VAL_3]] step %[[VAL_6]] {
// CHECK: %[[VAL_33:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_31]]] : memref<32xf32>
// CHECK: memref.store %[[VAL_33]], %[[VAL_12]]{{\[}}%[[VAL_31]]] : memref<32xf32>
// CHECK: }
// CHECK: %[[VAL_34:.*]] = bufferization.to_tensor %[[VAL_12]] : memref<32xf32>
// CHECK: return %[[VAL_34]] : tensor<32xf32>
// CHECK: }
func @add_ds(%arga: tensor<32xf32>, %argb: tensor<32xf32, #SV>, %argx: tensor<32xf32>) -> tensor<32xf32> {
%0 = linalg.generic #trait2
ins(%arga, %argb: tensor<32xf32>, tensor<32xf32, #SV>)
outs(%argx: tensor<32xf32>) {
^bb(%a: f32, %b: f32, %x: f32):
%0 = arith.addf %a, %b : f32
linalg.yield %0 : f32
} -> tensor<32xf32>
return %0 : tensor<32xf32>
}
// CHECK-LABEL: func @mul_ds(
// CHECK-SAME: %[[VAL_0:.*]]: tensor<32xf32>,
// CHECK-SAME: %[[VAL_1:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
// CHECK: %[[VAL_3:.*]] = arith.constant 0 : index
// CHECK: %[[VAL_4:.*]] = arith.constant 1 : index
// CHECK: %[[VAL_5:.*]] = bufferization.to_memref %[[VAL_0]] : memref<32xf32>
// CHECK: %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
// CHECK: %[[VAL_10:.*]] = memref.alloc() : memref<32xf32>
// CHECK: memref.copy %[[VAL_9]], %[[VAL_10]] : memref<32xf32> to memref<32xf32>
// CHECK: %[[VAL_11:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_3]]] : memref<?xindex>
// CHECK: %[[VAL_12:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_4]]] : memref<?xindex>
// CHECK: scf.for %[[VAL_13:.*]] = %[[VAL_11]] to %[[VAL_12]] step %[[VAL_4]] {
// CHECK: %[[VAL_14:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_13]]] : memref<?xindex>
// CHECK: %[[VAL_15:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_14]]] : memref<32xf32>
// CHECK: %[[VAL_16:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_13]]] : memref<?xf32>
// CHECK: %[[VAL_17:.*]] = arith.mulf %[[VAL_15]], %[[VAL_16]] : f32
// CHECK: memref.store %[[VAL_17]], %[[VAL_10]]{{\[}}%[[VAL_14]]] : memref<32xf32>
// CHECK: }
// CHECK: %[[VAL_18:.*]] = bufferization.to_tensor %[[VAL_10]] : memref<32xf32>
// CHECK: return %[[VAL_18]] : tensor<32xf32>
// CHECK: }
func @mul_ds(%arga: tensor<32xf32>, %argb: tensor<32xf32, #SV>, %argx: tensor<32xf32>) -> tensor<32xf32> {
%0 = linalg.generic #trait2
ins(%arga, %argb: tensor<32xf32>, tensor<32xf32, #SV>)
outs(%argx: tensor<32xf32>) {
^bb(%a: f32, %b: f32, %x: f32):
%0 = arith.mulf %a, %b : f32
linalg.yield %0 : f32
} -> tensor<32xf32>
return %0 : tensor<32xf32>
}
// CHECK-LABEL: func @add_sd(
// CHECK-SAME: %[[VAL_0:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_1:.*]]: tensor<32xf32>,
// CHECK-SAME: %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
// CHECK-DAG: %[[VAL_3:.*]] = arith.constant 32 : index
// CHECK-DAG: %[[VAL_4:.*]] = arith.constant 0 : index
// CHECK-DAG: %[[VAL_5:.*]] = arith.constant true
// CHECK-DAG: %[[VAL_6:.*]] = arith.constant 1 : index
// CHECK: %[[VAL_7:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_8:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_9:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_10:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xf32>
// CHECK: %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
// CHECK: %[[VAL_12:.*]] = memref.alloc() : memref<32xf32>
// CHECK: memref.copy %[[VAL_11]], %[[VAL_12]] : memref<32xf32> to memref<32xf32>
// CHECK: %[[VAL_13:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_4]]] : memref<?xindex>
// CHECK: %[[VAL_14:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_6]]] : memref<?xindex>
// CHECK: %[[VAL_15:.*]]:2 = scf.while (%[[VAL_16:.*]] = %[[VAL_13]], %[[VAL_17:.*]] = %[[VAL_4]]) : (index, index) -> (index, index) {
// CHECK: %[[VAL_18:.*]] = arith.cmpi ult, %[[VAL_16]], %[[VAL_14]] : index
// CHECK: scf.condition(%[[VAL_18]]) %[[VAL_16]], %[[VAL_17]] : index, index
// CHECK: } do {
// CHECK: ^bb0(%[[VAL_19:.*]]: index, %[[VAL_20:.*]]: index):
// CHECK: %[[VAL_21:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_19]]] : memref<?xindex>
// CHECK: %[[VAL_22:.*]] = arith.cmpi eq, %[[VAL_21]], %[[VAL_20]] : index
// CHECK: scf.if %[[VAL_22]] {
// CHECK: %[[VAL_23:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_19]]] : memref<?xf32>
// CHECK: %[[VAL_24:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_20]]] : memref<32xf32>
// CHECK: %[[VAL_25:.*]] = arith.addf %[[VAL_23]], %[[VAL_24]] : f32
// CHECK: memref.store %[[VAL_25]], %[[VAL_12]]{{\[}}%[[VAL_20]]] : memref<32xf32>
// CHECK: } else {
// CHECK: scf.if %[[VAL_5]] {
// CHECK: %[[VAL_26:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_20]]] : memref<32xf32>
// CHECK: memref.store %[[VAL_26]], %[[VAL_12]]{{\[}}%[[VAL_20]]] : memref<32xf32>
// CHECK: } else {
// CHECK: }
// CHECK: }
// CHECK: %[[VAL_27:.*]] = arith.cmpi eq, %[[VAL_21]], %[[VAL_20]] : index
// CHECK: %[[VAL_28:.*]] = arith.addi %[[VAL_19]], %[[VAL_6]] : index
// CHECK: %[[VAL_29:.*]] = select %[[VAL_27]], %[[VAL_28]], %[[VAL_19]] : index
// CHECK: %[[VAL_30:.*]] = arith.addi %[[VAL_20]], %[[VAL_6]] : index
// CHECK: scf.yield %[[VAL_29]], %[[VAL_30]] : index, index
// CHECK: }
// CHECK: scf.for %[[VAL_31:.*]] = %[[VAL_32:.*]]#1 to %[[VAL_3]] step %[[VAL_6]] {
// CHECK: %[[VAL_33:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_31]]] : memref<32xf32>
// CHECK: memref.store %[[VAL_33]], %[[VAL_12]]{{\[}}%[[VAL_31]]] : memref<32xf32>
// CHECK: }
// CHECK: %[[VAL_34:.*]] = bufferization.to_tensor %[[VAL_12]] : memref<32xf32>
// CHECK: return %[[VAL_34]] : tensor<32xf32>
// CHECK: }
func @add_sd(%arga: tensor<32xf32, #SV>, %argb: tensor<32xf32>, %argx: tensor<32xf32>) -> tensor<32xf32> {
%0 = linalg.generic #trait2
ins(%arga, %argb: tensor<32xf32, #SV>, tensor<32xf32>)
outs(%argx: tensor<32xf32>) {
^bb(%a: f32, %b: f32, %x: f32):
%0 = arith.addf %a, %b : f32
linalg.yield %0 : f32
} -> tensor<32xf32>
return %0 : tensor<32xf32>
}
// CHECK-LABEL: func @mul_sd(
// CHECK-SAME: %[[VAL_0:.*]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_1:.*]]: tensor<32xf32>,
// CHECK-SAME: %[[VAL_2:.*]]: tensor<32xf32>) -> tensor<32xf32> {
// CHECK: %[[VAL_3:.*]] = arith.constant 0 : index
// CHECK: %[[VAL_4:.*]] = arith.constant 1 : index
// CHECK: %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_1]] : memref<32xf32>
// CHECK: %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
// CHECK: %[[VAL_10:.*]] = memref.alloc() : memref<32xf32>
// CHECK: memref.copy %[[VAL_9]], %[[VAL_10]] : memref<32xf32> to memref<32xf32>
// CHECK: %[[VAL_11:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
// CHECK: %[[VAL_12:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
// CHECK: scf.for %[[VAL_13:.*]] = %[[VAL_11]] to %[[VAL_12]] step %[[VAL_4]] {
// CHECK: %[[VAL_14:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_13]]] : memref<?xindex>
// CHECK: %[[VAL_15:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_13]]] : memref<?xf32>
// CHECK: %[[VAL_16:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_14]]] : memref<32xf32>
// CHECK: %[[VAL_17:.*]] = arith.mulf %[[VAL_15]], %[[VAL_16]] : f32
// CHECK: memref.store %[[VAL_17]], %[[VAL_10]]{{\[}}%[[VAL_14]]] : memref<32xf32>
// CHECK: }
// CHECK: %[[VAL_18:.*]] = bufferization.to_tensor %[[VAL_10]] : memref<32xf32>
// CHECK: return %[[VAL_18]] : tensor<32xf32>
// CHECK: }
func @mul_sd(%arga: tensor<32xf32, #SV>, %argb: tensor<32xf32>, %argx: tensor<32xf32>) -> tensor<32xf32> {
%0 = linalg.generic #trait2
ins(%arga, %argb: tensor<32xf32, #SV>, tensor<32xf32>)
outs(%argx: tensor<32xf32>) {
^bb(%a: f32, %b: f32, %x: f32):
%0 = arith.mulf %a, %b : f32
linalg.yield %0 : f32
} -> tensor<32xf32>
return %0 : tensor<32xf32>
}
// CHECK-LABEL: func @add_ss(
// CHECK-SAME: %[[VAL_0:.*0]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_1:.*1]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_2:.*2]]: tensor<32xf32>) -> tensor<32xf32> {
// CHECK: %[[VAL_3:.*]] = arith.constant 0 : index
// CHECK: %[[VAL_4:.*]] = arith.constant 1 : index
// CHECK: %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
// CHECK: %[[VAL_12:.*]] = memref.alloc() : memref<32xf32>
// CHECK: memref.copy %[[VAL_11]], %[[VAL_12]] : memref<32xf32> to memref<32xf32>
// CHECK: %[[VAL_13:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
// CHECK: %[[VAL_14:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
// CHECK: %[[VAL_15:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_3]]] : memref<?xindex>
// CHECK: %[[VAL_16:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_4]]] : memref<?xindex>
// CHECK: %[[VAL_17:.*]]:2 = scf.while (%[[VAL_18:.*]] = %[[VAL_13]], %[[VAL_19:.*]] = %[[VAL_15]]) : (index, index) -> (index, index) {
// CHECK: %[[VAL_20:.*]] = arith.cmpi ult, %[[VAL_18]], %[[VAL_14]] : index
// CHECK: %[[VAL_21:.*]] = arith.cmpi ult, %[[VAL_19]], %[[VAL_16]] : index
// CHECK: %[[VAL_22:.*]] = arith.andi %[[VAL_20]], %[[VAL_21]] : i1
// CHECK: scf.condition(%[[VAL_22]]) %[[VAL_18]], %[[VAL_19]] : index, index
// CHECK: } do {
// CHECK: ^bb0(%[[VAL_23:.*]]: index, %[[VAL_24:.*]]: index):
// CHECK: %[[VAL_25:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_23]]] : memref<?xindex>
// CHECK: %[[VAL_26:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_24]]] : memref<?xindex>
// CHECK: %[[VAL_27:.*]] = arith.cmpi ult, %[[VAL_26]], %[[VAL_25]] : index
// CHECK: %[[VAL_28:.*]] = select %[[VAL_27]], %[[VAL_26]], %[[VAL_25]] : index
// CHECK: %[[VAL_29:.*]] = arith.cmpi eq, %[[VAL_25]], %[[VAL_28]] : index
// CHECK: %[[VAL_30:.*]] = arith.cmpi eq, %[[VAL_26]], %[[VAL_28]] : index
// CHECK: %[[VAL_31:.*]] = arith.andi %[[VAL_29]], %[[VAL_30]] : i1
// CHECK: scf.if %[[VAL_31]] {
// CHECK: %[[VAL_32:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_23]]] : memref<?xf32>
// CHECK: %[[VAL_33:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_24]]] : memref<?xf32>
// CHECK: %[[VAL_34:.*]] = arith.addf %[[VAL_32]], %[[VAL_33]] : f32
// CHECK: memref.store %[[VAL_34]], %[[VAL_12]]{{\[}}%[[VAL_28]]] : memref<32xf32>
// CHECK: } else {
// CHECK: %[[VAL_35:.*]] = arith.cmpi eq, %[[VAL_25]], %[[VAL_28]] : index
// CHECK: scf.if %[[VAL_35]] {
// CHECK: %[[VAL_36:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_23]]] : memref<?xf32>
// CHECK: memref.store %[[VAL_36]], %[[VAL_12]]{{\[}}%[[VAL_28]]] : memref<32xf32>
// CHECK: } else {
// CHECK: %[[VAL_37:.*]] = arith.cmpi eq, %[[VAL_26]], %[[VAL_28]] : index
// CHECK: scf.if %[[VAL_37]] {
// CHECK: %[[VAL_38:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_24]]] : memref<?xf32>
// CHECK: memref.store %[[VAL_38]], %[[VAL_12]]{{\[}}%[[VAL_28]]] : memref<32xf32>
// CHECK: } else {
// CHECK: }
// CHECK: }
// CHECK: }
// CHECK: %[[VAL_39:.*]] = arith.cmpi eq, %[[VAL_25]], %[[VAL_28]] : index
// CHECK: %[[VAL_40:.*]] = arith.addi %[[VAL_23]], %[[VAL_4]] : index
// CHECK: %[[VAL_41:.*]] = select %[[VAL_39]], %[[VAL_40]], %[[VAL_23]] : index
// CHECK: %[[VAL_42:.*]] = arith.cmpi eq, %[[VAL_26]], %[[VAL_28]] : index
// CHECK: %[[VAL_43:.*]] = arith.addi %[[VAL_24]], %[[VAL_4]] : index
// CHECK: %[[VAL_44:.*]] = select %[[VAL_42]], %[[VAL_43]], %[[VAL_24]] : index
// CHECK: scf.yield %[[VAL_41]], %[[VAL_44]] : index, index
// CHECK: }
// CHECK: scf.for %[[VAL_45:.*]] = %[[VAL_46:.*]]#0 to %[[VAL_14]] step %[[VAL_4]] {
// CHECK: %[[VAL_47:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_45]]] : memref<?xindex>
// CHECK: %[[VAL_48:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_45]]] : memref<?xf32>
// CHECK: memref.store %[[VAL_48]], %[[VAL_12]]{{\[}}%[[VAL_47]]] : memref<32xf32>
// CHECK: }
// CHECK: scf.for %[[VAL_49:.*]] = %[[VAL_50:.*]]#1 to %[[VAL_16]] step %[[VAL_4]] {
// CHECK: %[[VAL_51:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_49]]] : memref<?xindex>
// CHECK: %[[VAL_52:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_49]]] : memref<?xf32>
// CHECK: memref.store %[[VAL_52]], %[[VAL_12]]{{\[}}%[[VAL_51]]] : memref<32xf32>
// CHECK: }
// CHECK: %[[VAL_53:.*]] = bufferization.to_tensor %[[VAL_12]] : memref<32xf32>
// CHECK: return %[[VAL_53]] : tensor<32xf32>
// CHECK: }
func @add_ss(%arga: tensor<32xf32, #SV>, %argb: tensor<32xf32, #SV>, %argx: tensor<32xf32>) -> tensor<32xf32> {
%0 = linalg.generic #trait2
ins(%arga, %argb: tensor<32xf32, #SV>, tensor<32xf32, #SV>)
outs(%argx: tensor<32xf32>) {
^bb(%a: f32, %b: f32, %x: f32):
%0 = arith.addf %a, %b : f32
linalg.yield %0 : f32
} -> tensor<32xf32>
return %0 : tensor<32xf32>
}
// CHECK-LABEL: func @mul_ss(
// CHECK-SAME: %[[VAL_0:.*0]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_1:.*1]]: tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_2:.*2]]: tensor<32xf32>) -> tensor<32xf32> {
// CHECK: %[[VAL_3:.*]] = arith.constant 0 : index
// CHECK: %[[VAL_4:.*]] = arith.constant 1 : index
// CHECK: %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_3]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<32xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<32xf32>
// CHECK: %[[VAL_12:.*]] = memref.alloc() : memref<32xf32>
// CHECK: memref.copy %[[VAL_11]], %[[VAL_12]] : memref<32xf32> to memref<32xf32>
// CHECK: %[[VAL_13:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
// CHECK: %[[VAL_14:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
// CHECK: %[[VAL_15:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_3]]] : memref<?xindex>
// CHECK: %[[VAL_16:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_4]]] : memref<?xindex>
// CHECK: %[[VAL_17:.*]]:2 = scf.while (%[[VAL_18:.*]] = %[[VAL_13]], %[[VAL_19:.*]] = %[[VAL_15]]) : (index, index) -> (index, index) {
// CHECK: %[[VAL_20:.*]] = arith.cmpi ult, %[[VAL_18]], %[[VAL_14]] : index
// CHECK: %[[VAL_21:.*]] = arith.cmpi ult, %[[VAL_19]], %[[VAL_16]] : index
// CHECK: %[[VAL_22:.*]] = arith.andi %[[VAL_20]], %[[VAL_21]] : i1
// CHECK: scf.condition(%[[VAL_22]]) %[[VAL_18]], %[[VAL_19]] : index, index
// CHECK: } do {
// CHECK: ^bb0(%[[VAL_23:.*]]: index, %[[VAL_24:.*]]: index):
// CHECK: %[[VAL_25:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_23]]] : memref<?xindex>
// CHECK: %[[VAL_26:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_24]]] : memref<?xindex>
// CHECK: %[[VAL_27:.*]] = arith.cmpi ult, %[[VAL_26]], %[[VAL_25]] : index
// CHECK: %[[VAL_28:.*]] = select %[[VAL_27]], %[[VAL_26]], %[[VAL_25]] : index
// CHECK: %[[VAL_29:.*]] = arith.cmpi eq, %[[VAL_25]], %[[VAL_28]] : index
// CHECK: %[[VAL_30:.*]] = arith.cmpi eq, %[[VAL_26]], %[[VAL_28]] : index
// CHECK: %[[VAL_31:.*]] = arith.andi %[[VAL_29]], %[[VAL_30]] : i1
// CHECK: scf.if %[[VAL_31]] {
// CHECK: %[[VAL_32:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_23]]] : memref<?xf32>
// CHECK: %[[VAL_33:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_24]]] : memref<?xf32>
// CHECK: %[[VAL_34:.*]] = arith.mulf %[[VAL_32]], %[[VAL_33]] : f32
// CHECK: memref.store %[[VAL_34]], %[[VAL_12]]{{\[}}%[[VAL_28]]] : memref<32xf32>
// CHECK: } else {
// CHECK: }
// CHECK: %[[VAL_35:.*]] = arith.cmpi eq, %[[VAL_25]], %[[VAL_28]] : index
// CHECK: %[[VAL_36:.*]] = arith.addi %[[VAL_23]], %[[VAL_4]] : index
// CHECK: %[[VAL_37:.*]] = select %[[VAL_35]], %[[VAL_36]], %[[VAL_23]] : index
// CHECK: %[[VAL_38:.*]] = arith.cmpi eq, %[[VAL_26]], %[[VAL_28]] : index
// CHECK: %[[VAL_39:.*]] = arith.addi %[[VAL_24]], %[[VAL_4]] : index
// CHECK: %[[VAL_40:.*]] = select %[[VAL_38]], %[[VAL_39]], %[[VAL_24]] : index
// CHECK: scf.yield %[[VAL_37]], %[[VAL_40]] : index, index
// CHECK: }
// CHECK: %[[VAL_41:.*]] = bufferization.to_tensor %[[VAL_12]] : memref<32xf32>
// CHECK: return %[[VAL_41]] : tensor<32xf32>
// CHECK: }
func @mul_ss(%arga: tensor<32xf32, #SV>, %argb: tensor<32xf32, #SV>, %argx: tensor<32xf32>) -> tensor<32xf32> {
%0 = linalg.generic #trait2
ins(%arga, %argb: tensor<32xf32, #SV>, tensor<32xf32, #SV>)
outs(%argx: tensor<32xf32>) {
^bb(%a: f32, %b: f32, %x: f32):
%0 = arith.mulf %a, %b : f32
linalg.yield %0 : f32
} -> tensor<32xf32>
return %0 : tensor<32xf32>
}
// CHECK-LABEL: func @two_way_inv(
// CHECK-SAME: %[[VAL_0:.*0]]: tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_1:.*1]]: tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_2:.*2]]: f32,
// CHECK-SAME: %[[VAL_3:.*3]]: tensor<16xf32>) -> tensor<16xf32> {
// CHECK: %[[VAL_4:.*]] = arith.constant 0 : index
// CHECK: %[[VAL_5:.*]] = arith.constant 1 : index
// CHECK: %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_12:.*]] = bufferization.to_memref %[[VAL_3]] : memref<16xf32>
// CHECK: %[[VAL_13:.*]] = memref.alloc() : memref<16xf32>
// CHECK: memref.copy %[[VAL_12]], %[[VAL_13]] : memref<16xf32> to memref<16xf32>
// CHECK: %[[VAL_14:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_4]]] : memref<?xindex>
// CHECK: %[[VAL_15:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_5]]] : memref<?xindex>
// CHECK: %[[VAL_16:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_4]]] : memref<?xindex>
// CHECK: %[[VAL_17:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_5]]] : memref<?xindex>
// CHECK: %[[VAL_18:.*]]:2 = scf.while (%[[VAL_19:.*]] = %[[VAL_14]], %[[VAL_20:.*]] = %[[VAL_16]]) : (index, index) -> (index, index) {
// CHECK: %[[VAL_21:.*]] = arith.cmpi ult, %[[VAL_19]], %[[VAL_15]] : index
// CHECK: %[[VAL_22:.*]] = arith.cmpi ult, %[[VAL_20]], %[[VAL_17]] : index
// CHECK: %[[VAL_23:.*]] = arith.andi %[[VAL_21]], %[[VAL_22]] : i1
// CHECK: scf.condition(%[[VAL_23]]) %[[VAL_19]], %[[VAL_20]] : index, index
// CHECK: } do {
// CHECK: ^bb0(%[[VAL_24:.*]]: index, %[[VAL_25:.*]]: index):
// CHECK: %[[VAL_26:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_24]]] : memref<?xindex>
// CHECK: %[[VAL_27:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_25]]] : memref<?xindex>
// CHECK: %[[VAL_28:.*]] = arith.cmpi ult, %[[VAL_27]], %[[VAL_26]] : index
// CHECK: %[[VAL_29:.*]] = select %[[VAL_28]], %[[VAL_27]], %[[VAL_26]] : index
// CHECK: %[[VAL_30:.*]] = arith.cmpi eq, %[[VAL_26]], %[[VAL_29]] : index
// CHECK: %[[VAL_31:.*]] = arith.cmpi eq, %[[VAL_27]], %[[VAL_29]] : index
// CHECK: %[[VAL_32:.*]] = arith.andi %[[VAL_30]], %[[VAL_31]] : i1
// CHECK: scf.if %[[VAL_32]] {
// CHECK: %[[VAL_33:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_24]]] : memref<?xf32>
// CHECK: %[[VAL_34:.*]] = arith.mulf %[[VAL_33]], %[[VAL_2]] : f32
// CHECK: %[[VAL_35:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_25]]] : memref<?xf32>
// CHECK: %[[VAL_36:.*]] = arith.mulf %[[VAL_35]], %[[VAL_2]] : f32
// CHECK: %[[VAL_37:.*]] = arith.addf %[[VAL_34]], %[[VAL_36]] : f32
// CHECK: memref.store %[[VAL_37]], %[[VAL_13]]{{\[}}%[[VAL_29]]] : memref<16xf32>
// CHECK: } else {
// CHECK: %[[VAL_38:.*]] = arith.cmpi eq, %[[VAL_26]], %[[VAL_29]] : index
// CHECK: scf.if %[[VAL_38]] {
// CHECK: %[[VAL_39:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_24]]] : memref<?xf32>
// CHECK: %[[VAL_40:.*]] = arith.mulf %[[VAL_39]], %[[VAL_2]] : f32
// CHECK: memref.store %[[VAL_40]], %[[VAL_13]]{{\[}}%[[VAL_29]]] : memref<16xf32>
// CHECK: } else {
// CHECK: %[[VAL_41:.*]] = arith.cmpi eq, %[[VAL_27]], %[[VAL_29]] : index
// CHECK: scf.if %[[VAL_41]] {
// CHECK: %[[VAL_42:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_25]]] : memref<?xf32>
// CHECK: %[[VAL_43:.*]] = arith.mulf %[[VAL_42]], %[[VAL_2]] : f32
// CHECK: memref.store %[[VAL_43]], %[[VAL_13]]{{\[}}%[[VAL_29]]] : memref<16xf32>
// CHECK: } else {
// CHECK: }
// CHECK: }
// CHECK: }
// CHECK: %[[VAL_44:.*]] = arith.cmpi eq, %[[VAL_26]], %[[VAL_29]] : index
// CHECK: %[[VAL_45:.*]] = arith.addi %[[VAL_24]], %[[VAL_5]] : index
// CHECK: %[[VAL_46:.*]] = select %[[VAL_44]], %[[VAL_45]], %[[VAL_24]] : index
// CHECK: %[[VAL_47:.*]] = arith.cmpi eq, %[[VAL_27]], %[[VAL_29]] : index
// CHECK: %[[VAL_48:.*]] = arith.addi %[[VAL_25]], %[[VAL_5]] : index
// CHECK: %[[VAL_49:.*]] = select %[[VAL_47]], %[[VAL_48]], %[[VAL_25]] : index
// CHECK: scf.yield %[[VAL_46]], %[[VAL_49]] : index, index
// CHECK: }
// CHECK: scf.for %[[VAL_50:.*]] = %[[VAL_51:.*]]#0 to %[[VAL_15]] step %[[VAL_5]] {
// CHECK: %[[VAL_52:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_50]]] : memref<?xindex>
// CHECK: %[[VAL_53:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_50]]] : memref<?xf32>
// CHECK: %[[VAL_54:.*]] = arith.mulf %[[VAL_53]], %[[VAL_2]] : f32
// CHECK: memref.store %[[VAL_54]], %[[VAL_13]]{{\[}}%[[VAL_52]]] : memref<16xf32>
// CHECK: }
// CHECK: scf.for %[[VAL_55:.*]] = %[[VAL_56:.*]]#1 to %[[VAL_17]] step %[[VAL_5]] {
// CHECK: %[[VAL_57:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_55]]] : memref<?xindex>
// CHECK: %[[VAL_58:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_55]]] : memref<?xf32>
// CHECK: %[[VAL_59:.*]] = arith.mulf %[[VAL_58]], %[[VAL_2]] : f32
// CHECK: memref.store %[[VAL_59]], %[[VAL_13]]{{\[}}%[[VAL_57]]] : memref<16xf32>
// CHECK: }
// CHECK: %[[VAL_60:.*]] = bufferization.to_tensor %[[VAL_13]] : memref<16xf32>
// CHECK: return %[[VAL_60]] : tensor<16xf32>
// CHECK: }
func @two_way_inv(%arga: tensor<16xf32, #SV>, %argb: tensor<16xf32, #SV>, %argc: f32, %argx: tensor<16xf32>) -> tensor<16xf32> {
// Kernel "x(i) = a(i) * c + b(i) * c".
%0 = linalg.generic #trait2
ins(%arga, %argb: tensor<16xf32, #SV>, tensor<16xf32, #SV>)
outs(%argx: tensor<16xf32>) {
^bb(%a: f32, %b: f32, %x: f32):
%0 = arith.mulf %a, %argc : f32
%1 = arith.mulf %b, %argc : f32
%2 = arith.addf %0, %1 : f32
linalg.yield %2 : f32
} -> tensor<16xf32>
return %0 : tensor<16xf32>
}
// CHECK-LABEL: func @two_way_inv_alt(
// CHECK-SAME: %[[VAL_0:.*0]]: tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_1:.*1]]: tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_2:.*2]]: f32,
// CHECK-SAME: %[[VAL_3:.*3]]: tensor<16xf32>) -> tensor<16xf32> {
// CHECK: %[[VAL_4:.*]] = arith.constant 0 : index
// CHECK: %[[VAL_5:.*]] = arith.constant 1 : index
// CHECK: %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_12:.*]] = bufferization.to_memref %[[VAL_3]] : memref<16xf32>
// CHECK: %[[VAL_13:.*]] = memref.alloc() : memref<16xf32>
// CHECK: memref.copy %[[VAL_12]], %[[VAL_13]] : memref<16xf32> to memref<16xf32>
// CHECK: %[[VAL_14:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_4]]] : memref<?xindex>
// CHECK: %[[VAL_15:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_5]]] : memref<?xindex>
// CHECK: %[[VAL_16:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_4]]] : memref<?xindex>
// CHECK: %[[VAL_17:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_5]]] : memref<?xindex>
// CHECK: %[[VAL_18:.*]]:2 = scf.while (%[[VAL_19:.*]] = %[[VAL_14]], %[[VAL_20:.*]] = %[[VAL_16]]) : (index, index) -> (index, index) {
// CHECK: %[[VAL_21:.*]] = arith.cmpi ult, %[[VAL_19]], %[[VAL_15]] : index
// CHECK: %[[VAL_22:.*]] = arith.cmpi ult, %[[VAL_20]], %[[VAL_17]] : index
// CHECK: %[[VAL_23:.*]] = arith.andi %[[VAL_21]], %[[VAL_22]] : i1
// CHECK: scf.condition(%[[VAL_23]]) %[[VAL_19]], %[[VAL_20]] : index, index
// CHECK: } do {
// CHECK: ^bb0(%[[VAL_24:.*]]: index, %[[VAL_25:.*]]: index):
// CHECK: %[[VAL_26:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_24]]] : memref<?xindex>
// CHECK: %[[VAL_27:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_25]]] : memref<?xindex>
// CHECK: %[[VAL_28:.*]] = arith.cmpi ult, %[[VAL_27]], %[[VAL_26]] : index
// CHECK: %[[VAL_29:.*]] = select %[[VAL_28]], %[[VAL_27]], %[[VAL_26]] : index
// CHECK: %[[VAL_30:.*]] = arith.cmpi eq, %[[VAL_26]], %[[VAL_29]] : index
// CHECK: %[[VAL_31:.*]] = arith.cmpi eq, %[[VAL_27]], %[[VAL_29]] : index
// CHECK: %[[VAL_32:.*]] = arith.andi %[[VAL_30]], %[[VAL_31]] : i1
// CHECK: scf.if %[[VAL_32]] {
// CHECK: %[[VAL_33:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_24]]] : memref<?xf32>
// CHECK: %[[VAL_34:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_25]]] : memref<?xf32>
// CHECK: %[[VAL_35:.*]] = arith.addf %[[VAL_33]], %[[VAL_34]] : f32
// CHECK: %[[VAL_36:.*]] = arith.mulf %[[VAL_35]], %[[VAL_2]] : f32
// CHECK: memref.store %[[VAL_36]], %[[VAL_13]]{{\[}}%[[VAL_29]]] : memref<16xf32>
// CHECK: } else {
// CHECK: %[[VAL_37:.*]] = arith.cmpi eq, %[[VAL_26]], %[[VAL_29]] : index
// CHECK: scf.if %[[VAL_37]] {
// CHECK: %[[VAL_38:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_24]]] : memref<?xf32>
// CHECK: %[[VAL_39:.*]] = arith.mulf %[[VAL_38]], %[[VAL_2]] : f32
// CHECK: memref.store %[[VAL_39]], %[[VAL_13]]{{\[}}%[[VAL_29]]] : memref<16xf32>
// CHECK: } else {
// CHECK: %[[VAL_40:.*]] = arith.cmpi eq, %[[VAL_27]], %[[VAL_29]] : index
// CHECK: scf.if %[[VAL_40]] {
// CHECK: %[[VAL_41:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_25]]] : memref<?xf32>
// CHECK: %[[VAL_42:.*]] = arith.mulf %[[VAL_41]], %[[VAL_2]] : f32
// CHECK: memref.store %[[VAL_42]], %[[VAL_13]]{{\[}}%[[VAL_29]]] : memref<16xf32>
// CHECK: } else {
// CHECK: }
// CHECK: }
// CHECK: }
// CHECK: %[[VAL_43:.*]] = arith.cmpi eq, %[[VAL_26]], %[[VAL_29]] : index
// CHECK: %[[VAL_44:.*]] = arith.addi %[[VAL_24]], %[[VAL_5]] : index
// CHECK: %[[VAL_45:.*]] = select %[[VAL_43]], %[[VAL_44]], %[[VAL_24]] : index
// CHECK: %[[VAL_46:.*]] = arith.cmpi eq, %[[VAL_27]], %[[VAL_29]] : index
// CHECK: %[[VAL_47:.*]] = arith.addi %[[VAL_25]], %[[VAL_5]] : index
// CHECK: %[[VAL_48:.*]] = select %[[VAL_46]], %[[VAL_47]], %[[VAL_25]] : index
// CHECK: scf.yield %[[VAL_45]], %[[VAL_48]] : index, index
// CHECK: }
// CHECK: scf.for %[[VAL_49:.*]] = %[[VAL_50:.*]]#0 to %[[VAL_15]] step %[[VAL_5]] {
// CHECK: %[[VAL_51:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_49]]] : memref<?xindex>
// CHECK: %[[VAL_52:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_49]]] : memref<?xf32>
// CHECK: %[[VAL_53:.*]] = arith.mulf %[[VAL_52]], %[[VAL_2]] : f32
// CHECK: memref.store %[[VAL_53]], %[[VAL_13]]{{\[}}%[[VAL_51]]] : memref<16xf32>
// CHECK: }
// CHECK: scf.for %[[VAL_54:.*]] = %[[VAL_55:.*]]#1 to %[[VAL_17]] step %[[VAL_5]] {
// CHECK: %[[VAL_56:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_54]]] : memref<?xindex>
// CHECK: %[[VAL_57:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_54]]] : memref<?xf32>
// CHECK: %[[VAL_58:.*]] = arith.mulf %[[VAL_57]], %[[VAL_2]] : f32
// CHECK: memref.store %[[VAL_58]], %[[VAL_13]]{{\[}}%[[VAL_56]]] : memref<16xf32>
// CHECK: }
// CHECK: %[[VAL_59:.*]] = bufferization.to_tensor %[[VAL_13]] : memref<16xf32>
// CHECK: return %[[VAL_59]] : tensor<16xf32>
// CHECK: }
func @two_way_inv_alt(%arga: tensor<16xf32, #SV>,
%argb: tensor<16xf32, #SV>, %argc: f32, %argx: tensor<16xf32>) -> tensor<16xf32> {
// Same kernel, but now expressed as "x(i) = (a(i) + b(i)) * c".
%0 = linalg.generic #trait2
ins(%arga, %argb: tensor<16xf32, #SV>, tensor<16xf32, #SV>)
outs(%argx: tensor<16xf32>) {
^bb(%a: f32, %b: f32, %x: f32):
%0 = arith.addf %a, %b : f32
%1 = arith.mulf %0, %argc : f32
linalg.yield %1 : f32
} -> tensor<16xf32>
return %0 : tensor<16xf32>
}
#trait_sum_reduction = {
indexing_maps = [
affine_map<(i) -> (i)>, // a
affine_map<(i) -> ()> // x (scalar out)
],
iterator_types = ["reduction"],
doc = "x += SUM_i a(i)"
}
// CHECK-LABEL: func @sum_reduction(
// CHECK-SAME: %[[VAL_0:.*]]: tensor<?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_1:.*]]: tensor<f32>) -> tensor<f32> {
// CHECK-DAG: %[[VAL_2:.*]] = arith.constant 0 : index
// CHECK-DAG: %[[VAL_3:.*]] = arith.constant 1 : index
// CHECK: %[[VAL_4:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_2]] : tensor<?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_5:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<?xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_6:.*]] = bufferization.to_memref %[[VAL_1]] : memref<f32>
// CHECK: %[[VAL_7:.*]] = memref.alloc() : memref<f32>
// CHECK: memref.copy %[[VAL_6]], %[[VAL_7]] : memref<f32> to memref<f32>
// CHECK-DAG: %[[VAL_8:.*]] = memref.load %[[VAL_4]]{{\[}}%[[VAL_2]]] : memref<?xindex>
// CHECK-DAG: %[[VAL_9:.*]] = memref.load %[[VAL_4]]{{\[}}%[[VAL_3]]] : memref<?xindex>
// CHECK-DAG: %[[VAL_10:.*]] = memref.load %[[VAL_7]][] : memref<f32>
// CHECK: %[[VAL_11:.*]] = scf.for %[[VAL_12:.*]] = %[[VAL_8]] to %[[VAL_9]] step %[[VAL_3]] iter_args(%[[VAL_13:.*]] = %[[VAL_10]]) -> (f32) {
// CHECK: %[[VAL_14:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_12]]] : memref<?xf32>
// CHECK: %[[VAL_15:.*]] = arith.addf %[[VAL_13]], %[[VAL_14]] : f32
// CHECK: scf.yield %[[VAL_15]] : f32
// CHECK: }
// CHECK: memref.store %[[VAL_11]], %[[VAL_7]][] : memref<f32>
// CHECK: %[[VAL_17:.*]] = bufferization.to_tensor %[[VAL_7]] : memref<f32>
// CHECK: return %[[VAL_17]] : tensor<f32>
// CHECK: }
func @sum_reduction(%arga: tensor<?xf32, #SV>, %argx: tensor<f32>) -> tensor<f32> {
%0 = linalg.generic #trait_sum_reduction
ins(%arga: tensor<?xf32, #SV>)
outs(%argx: tensor<f32>) {
^bb(%a: f32, %x: f32):
%0 = arith.addf %x, %a : f32
linalg.yield %0 : f32
} -> tensor<f32>
return %0 : tensor<f32>
}
#trait_sum_reduction2 = {
indexing_maps = [
affine_map<(i) -> (i)>, // a
affine_map<(i) -> (i)>, // b
affine_map<(i)-> ()> // x (scalar out)
],
iterator_types = ["reduction"],
doc = "x += SUM_i a(i) + b(i)"
}
// CHECK-LABEL: func @sum_reduction_ss(
// CHECK-SAME: %[[VAL_0:.*0]]: tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_1:.*1]]: tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_2:.*2]]: tensor<f32>) -> tensor<f32> {
// CHECK-DAG: %[[VAL_3:.*]] = arith.constant 0 : index
// CHECK-DAG: %[[VAL_4:.*]] = arith.constant 1 : index
// CHECK: %[[VAL_5:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_3]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_6:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_3]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_7:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_8:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_3]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_9:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_3]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_10:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_11:.*]] = bufferization.to_memref %[[VAL_2]] : memref<f32>
// CHECK: %[[VAL_12:.*]] = memref.alloc() : memref<f32>
// CHECK: memref.copy %[[VAL_11]], %[[VAL_12]] : memref<f32> to memref<f32>
// CHECK: %[[VAL_13:.*]] = memref.load %[[VAL_12]][] : memref<f32>
// CHECK: %[[VAL_14:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_3]]] : memref<?xindex>
// CHECK: %[[VAL_15:.*]] = memref.load %[[VAL_5]]{{\[}}%[[VAL_4]]] : memref<?xindex>
// CHECK: %[[VAL_16:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_3]]] : memref<?xindex>
// CHECK: %[[VAL_17:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_4]]] : memref<?xindex>
// CHECK: %[[VAL_18:.*]]:3 = scf.while (%[[VAL_19:.*]] = %[[VAL_14]], %[[VAL_20:.*]] = %[[VAL_16]], %[[VAL_21:.*]] = %[[VAL_13]]) : (index, index, f32) -> (index, index, f32) {
// CHECK: %[[VAL_22:.*]] = arith.cmpi ult, %[[VAL_19]], %[[VAL_15]] : index
// CHECK: %[[VAL_23:.*]] = arith.cmpi ult, %[[VAL_20]], %[[VAL_17]] : index
// CHECK: %[[VAL_24:.*]] = arith.andi %[[VAL_22]], %[[VAL_23]] : i1
// CHECK: scf.condition(%[[VAL_24]]) %[[VAL_19]], %[[VAL_20]], %[[VAL_21]] : index, index, f32
// CHECK: } do {
// CHECK: ^bb0(%[[VAL_25:.*]]: index, %[[VAL_26:.*]]: index, %[[VAL_27:.*]]: f32):
// CHECK: %[[VAL_28:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_25]]] : memref<?xindex>
// CHECK: %[[VAL_29:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_26]]] : memref<?xindex>
// CHECK: %[[VAL_30:.*]] = arith.cmpi ult, %[[VAL_29]], %[[VAL_28]] : index
// CHECK: %[[VAL_31:.*]] = select %[[VAL_30]], %[[VAL_29]], %[[VAL_28]] : index
// CHECK: %[[VAL_32:.*]] = arith.cmpi eq, %[[VAL_28]], %[[VAL_31]] : index
// CHECK: %[[VAL_33:.*]] = arith.cmpi eq, %[[VAL_29]], %[[VAL_31]] : index
// CHECK: %[[VAL_34:.*]] = arith.andi %[[VAL_32]], %[[VAL_33]] : i1
// CHECK: %[[VAL_35:.*]] = scf.if %[[VAL_34]] -> (f32) {
// CHECK: %[[VAL_36:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_25]]] : memref<?xf32>
// CHECK: %[[VAL_37:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_26]]] : memref<?xf32>
// CHECK: %[[VAL_38:.*]] = arith.addf %[[VAL_36]], %[[VAL_37]] : f32
// CHECK: %[[VAL_39:.*]] = arith.addf %[[VAL_27]], %[[VAL_38]] : f32
// CHECK: scf.yield %[[VAL_39]] : f32
// CHECK: } else {
// CHECK: %[[VAL_40:.*]] = arith.cmpi eq, %[[VAL_28]], %[[VAL_31]] : index
// CHECK: %[[VAL_41:.*]] = scf.if %[[VAL_40]] -> (f32) {
// CHECK: %[[VAL_42:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_25]]] : memref<?xf32>
// CHECK: %[[VAL_43:.*]] = arith.addf %[[VAL_27]], %[[VAL_42]] : f32
// CHECK: scf.yield %[[VAL_43]] : f32
// CHECK: } else {
// CHECK: %[[VAL_44:.*]] = arith.cmpi eq, %[[VAL_29]], %[[VAL_31]] : index
// CHECK: %[[VAL_45:.*]] = scf.if %[[VAL_44]] -> (f32) {
// CHECK: %[[VAL_46:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_26]]] : memref<?xf32>
// CHECK: %[[VAL_47:.*]] = arith.addf %[[VAL_27]], %[[VAL_46]] : f32
// CHECK: scf.yield %[[VAL_47]] : f32
// CHECK: } else {
// CHECK: scf.yield %[[VAL_27]] : f32
// CHECK: }
// CHECK: scf.yield %[[VAL_48:.*]] : f32
// CHECK: }
// CHECK: scf.yield %[[VAL_49:.*]] : f32
// CHECK: }
// CHECK: %[[VAL_50:.*]] = arith.cmpi eq, %[[VAL_28]], %[[VAL_31]] : index
// CHECK: %[[VAL_51:.*]] = arith.addi %[[VAL_25]], %[[VAL_4]] : index
// CHECK: %[[VAL_52:.*]] = select %[[VAL_50]], %[[VAL_51]], %[[VAL_25]] : index
// CHECK: %[[VAL_53:.*]] = arith.cmpi eq, %[[VAL_29]], %[[VAL_31]] : index
// CHECK: %[[VAL_54:.*]] = arith.addi %[[VAL_26]], %[[VAL_4]] : index
// CHECK: %[[VAL_55:.*]] = select %[[VAL_53]], %[[VAL_54]], %[[VAL_26]] : index
// CHECK: scf.yield %[[VAL_52]], %[[VAL_55]], %[[VAL_56:.*]] : index, index, f32
// CHECK: }
// CHECK: %[[VAL_57:.*]] = scf.for %[[VAL_58:.*]] = %[[VAL_59:.*]]#0 to %[[VAL_15]] step %[[VAL_4]] iter_args(%[[VAL_60:.*]] = %[[VAL_59]]#2) -> (f32) {
// CHECK: %[[VAL_61:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_58]]] : memref<?xf32>
// CHECK: %[[VAL_62:.*]] = arith.addf %[[VAL_60]], %[[VAL_61]] : f32
// CHECK: scf.yield %[[VAL_62]] : f32
// CHECK: }
// CHECK: %[[VAL_63:.*]] = scf.for %[[VAL_64:.*]] = %[[VAL_65:.*]]#1 to %[[VAL_17]] step %[[VAL_4]] iter_args(%[[VAL_66:.*]] = %[[VAL_67:.*]]) -> (f32) {
// CHECK: %[[VAL_68:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_64]]] : memref<?xf32>
// CHECK: %[[VAL_69:.*]] = arith.addf %[[VAL_66]], %[[VAL_68]] : f32
// CHECK: scf.yield %[[VAL_69]] : f32
// CHECK: }
// CHECK: memref.store %[[VAL_70:.*]], %[[VAL_12]][] : memref<f32>
// CHECK: %[[VAL_71:.*]] = bufferization.to_tensor %[[VAL_12]] : memref<f32>
// CHECK: return %[[VAL_71]] : tensor<f32>
// CHECK: }
func @sum_reduction_ss(%arga: tensor<16xf32, #SV>,
%argb: tensor<16xf32, #SV>,
%argx: tensor<f32>) -> tensor<f32> {
// Just for testing. This case would be better expressed
// as two separate reductions kernels.
%0 = linalg.generic #trait_sum_reduction2
ins(%arga, %argb: tensor<16xf32, #SV>, tensor<16xf32, #SV>)
outs(%argx: tensor<f32>) {
^bb(%a: f32, %b: f32, %x: f32):
%0 = arith.addf %a, %b : f32
%1 = arith.addf %x, %0 : f32
linalg.yield %1 : f32
} -> tensor<f32>
return %0 : tensor<f32>
}
#trait_sum_reduction_inv = {
indexing_maps = [
affine_map<(i) -> (i)>, // a
affine_map<(i) -> ()>, // b
affine_map<(i) -> (i)>, // c
affine_map<(i) -> ()> // x (out)
],
iterator_types = ["reduction"],
doc = "x += SUM_i a(i) * b + c(i)"
}
// CHECK-LABEL: func @sum_reduction_inv(
// CHECK-SAME: %[[VAL_0:.*0]]: tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_1:.*1]]: tensor<f32>,
// CHECK-SAME: %[[VAL_2:.*2]]: tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_3:.*3]]: tensor<f32>) -> tensor<f32> {
// CHECK-DAG: %[[VAL_4:.*]] = arith.constant 0 : index
// CHECK-DAG: %[[VAL_5:.*]] = arith.constant 1 : index
// CHECK: %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_9:.*]] = bufferization.to_memref %[[VAL_1]] : memref<f32>
// CHECK: %[[VAL_10:.*]] = sparse_tensor.pointers %[[VAL_2]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_11:.*]] = sparse_tensor.indices %[[VAL_2]], %[[VAL_4]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_12:.*]] = sparse_tensor.values %[[VAL_2]] : tensor<16xf32, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf32>
// CHECK: %[[VAL_13:.*]] = bufferization.to_memref %[[VAL_3]] : memref<f32>
// CHECK: %[[VAL_14:.*]] = memref.alloc() : memref<f32>
// CHECK: memref.copy %[[VAL_13]], %[[VAL_14]] : memref<f32> to memref<f32>
// CHECK: %[[VAL_15:.*]] = memref.load %[[VAL_14]][] : memref<f32>
// CHECK: %[[VAL_16:.*]] = memref.load %[[VAL_9]][] : memref<f32>
// CHECK: %[[VAL_17:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_4]]] : memref<?xindex>
// CHECK: %[[VAL_18:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_5]]] : memref<?xindex>
// CHECK: %[[VAL_19:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_4]]] : memref<?xindex>
// CHECK: %[[VAL_20:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_5]]] : memref<?xindex>
// CHECK: %[[VAL_21:.*]]:3 = scf.while (%[[VAL_22:.*]] = %[[VAL_17]], %[[VAL_23:.*]] = %[[VAL_19]], %[[VAL_24:.*]] = %[[VAL_15]]) : (index, index, f32) -> (index, index, f32) {
// CHECK: %[[VAL_25:.*]] = arith.cmpi ult, %[[VAL_22]], %[[VAL_18]] : index
// CHECK: %[[VAL_26:.*]] = arith.cmpi ult, %[[VAL_23]], %[[VAL_20]] : index
// CHECK: %[[VAL_27:.*]] = arith.andi %[[VAL_25]], %[[VAL_26]] : i1
// CHECK: scf.condition(%[[VAL_27]]) %[[VAL_22]], %[[VAL_23]], %[[VAL_24]] : index, index, f32
// CHECK: } do {
// CHECK: ^bb0(%[[VAL_28:.*]]: index, %[[VAL_29:.*]]: index, %[[VAL_30:.*]]: f32):
// CHECK: %[[VAL_31:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_28]]] : memref<?xindex>
// CHECK: %[[VAL_32:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_29]]] : memref<?xindex>
// CHECK: %[[VAL_33:.*]] = arith.cmpi ult, %[[VAL_32]], %[[VAL_31]] : index
// CHECK: %[[VAL_34:.*]] = select %[[VAL_33]], %[[VAL_32]], %[[VAL_31]] : index
// CHECK: %[[VAL_35:.*]] = arith.cmpi eq, %[[VAL_31]], %[[VAL_34]] : index
// CHECK: %[[VAL_36:.*]] = arith.cmpi eq, %[[VAL_32]], %[[VAL_34]] : index
// CHECK: %[[VAL_37:.*]] = arith.andi %[[VAL_35]], %[[VAL_36]] : i1
// CHECK: %[[VAL_38:.*]] = scf.if %[[VAL_37]] -> (f32) {
// CHECK: %[[VAL_39:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_28]]] : memref<?xf32>
// CHECK: %[[VAL_40:.*]] = arith.mulf %[[VAL_39]], %[[VAL_16]] : f32
// CHECK: %[[VAL_41:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_29]]] : memref<?xf32>
// CHECK: %[[VAL_42:.*]] = arith.addf %[[VAL_40]], %[[VAL_41]] : f32
// CHECK: %[[VAL_43:.*]] = arith.addf %[[VAL_30]], %[[VAL_42]] : f32
// CHECK: scf.yield %[[VAL_43]] : f32
// CHECK: } else {
// CHECK: %[[VAL_44:.*]] = arith.cmpi eq, %[[VAL_31]], %[[VAL_34]] : index
// CHECK: %[[VAL_45:.*]] = scf.if %[[VAL_44]] -> (f32) {
// CHECK: %[[VAL_46:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_28]]] : memref<?xf32>
// CHECK: %[[VAL_47:.*]] = arith.mulf %[[VAL_46]], %[[VAL_16]] : f32
// CHECK: %[[VAL_48:.*]] = arith.addf %[[VAL_30]], %[[VAL_47]] : f32
// CHECK: scf.yield %[[VAL_48]] : f32
// CHECK: } else {
// CHECK: %[[VAL_49:.*]] = arith.cmpi eq, %[[VAL_32]], %[[VAL_34]] : index
// CHECK: %[[VAL_50:.*]] = scf.if %[[VAL_49]] -> (f32) {
// CHECK: %[[VAL_51:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_29]]] : memref<?xf32>
// CHECK: %[[VAL_52:.*]] = arith.addf %[[VAL_30]], %[[VAL_51]] : f32
// CHECK: scf.yield %[[VAL_52]] : f32
// CHECK: } else {
// CHECK: scf.yield %[[VAL_30]] : f32
// CHECK: }
// CHECK: scf.yield %[[VAL_53:.*]] : f32
// CHECK: }
// CHECK: scf.yield %[[VAL_54:.*]] : f32
// CHECK: }
// CHECK: %[[VAL_55:.*]] = arith.cmpi eq, %[[VAL_31]], %[[VAL_34]] : index
// CHECK: %[[VAL_56:.*]] = arith.addi %[[VAL_28]], %[[VAL_5]] : index
// CHECK: %[[VAL_57:.*]] = select %[[VAL_55]], %[[VAL_56]], %[[VAL_28]] : index
// CHECK: %[[VAL_58:.*]] = arith.cmpi eq, %[[VAL_32]], %[[VAL_34]] : index
// CHECK: %[[VAL_59:.*]] = arith.addi %[[VAL_29]], %[[VAL_5]] : index
// CHECK: %[[VAL_60:.*]] = select %[[VAL_58]], %[[VAL_59]], %[[VAL_29]] : index
// CHECK: scf.yield %[[VAL_57]], %[[VAL_60]], %[[VAL_61:.*]] : index, index, f32
// CHECK: }
// CHECK: %[[VAL_62:.*]] = scf.for %[[VAL_63:.*]] = %[[VAL_64:.*]]#0 to %[[VAL_18]] step %[[VAL_5]] iter_args(%[[VAL_65:.*]] = %[[VAL_64]]#2) -> (f32) {
// CHECK: %[[VAL_66:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_63]]] : memref<?xf32>
// CHECK: %[[VAL_67:.*]] = arith.mulf %[[VAL_66]], %[[VAL_16]] : f32
// CHECK: %[[VAL_68:.*]] = arith.addf %[[VAL_65]], %[[VAL_67]] : f32
// CHECK: scf.yield %[[VAL_68]] : f32
// CHECK: }
// CHECK: %[[VAL_69:.*]] = scf.for %[[VAL_70:.*]] = %[[VAL_71:.*]]#1 to %[[VAL_20]] step %[[VAL_5]] iter_args(%[[VAL_72:.*]] = %[[VAL_73:.*]]) -> (f32) {
// CHECK: %[[VAL_74:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_70]]] : memref<?xf32>
// CHECK: %[[VAL_75:.*]] = arith.addf %[[VAL_72]], %[[VAL_74]] : f32
// CHECK: scf.yield %[[VAL_75]] : f32
// CHECK: }
// CHECK: memref.store %[[VAL_76:.*]], %[[VAL_14]][] : memref<f32>
// CHECK: %[[VAL_77:.*]] = bufferization.to_tensor %[[VAL_14]] : memref<f32>
// CHECK: return %[[VAL_77]] : tensor<f32>
// CHECK: }
func @sum_reduction_inv(%arga: tensor<16xf32, #SV>,
%argb: tensor<f32>,
%argc: tensor<16xf32, #SV>,
%argx: tensor<f32>) -> tensor<f32> {
// Just for testing. This case would be better expressed
// as two separate reductions kernels.
%0 = linalg.generic #trait_sum_reduction_inv
ins(%arga, %argb, %argc : tensor<16xf32, #SV>, tensor<f32>, tensor<16xf32, #SV>)
outs(%argx: tensor<f32>) {
^bb(%a: f32, %b: f32, %c: f32, %x: f32):
%0 = arith.mulf %a, %b : f32
%1 = arith.addf %0, %c : f32
%2 = arith.addf %x, %1 : f32
linalg.yield %2 : f32
} -> tensor<f32>
return %0 : tensor<f32>
}
#trait_four_tensors = {
indexing_maps = [
affine_map<(i) -> (i)>, // A
affine_map<(i) -> (i)>, // B
affine_map<(i) -> (i)>, // C
affine_map<(i) -> (i)>, // D
affine_map<(i) -> (i)> // X (out)
],
iterator_types = ["parallel"],
doc = "X(i) = A(i) + B(i) + C(i) + D(i)"
}
// CHECK-LABEL: func @four_tensors_op(
// CHECK-SAME: %[[VAL_0:.*0]]: tensor<?xf64>,
// CHECK-SAME: %[[VAL_1:.*1]]: tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_2:.*2]]: tensor<?xf64>,
// CHECK-SAME: %[[VAL_3:.*3]]: tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_4:.*]]: tensor<?xf64>) -> tensor<?xf64> {
// CHECK-DAG: %[[VAL_5:.*]] = arith.constant 0 : index
// CHECK-DAG: %[[VAL_6:.*]] = arith.constant true
// CHECK-DAG: %[[VAL_7:.*]] = arith.constant 1 : index
// CHECK: %[[VAL_8:.*]] = bufferization.to_memref %[[VAL_0]] : memref<?xf64>
// CHECK: %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_5]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_5]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf64>
// CHECK: %[[VAL_12:.*]] = bufferization.to_memref %[[VAL_2]] : memref<?xf64>
// CHECK: %[[VAL_13:.*]] = sparse_tensor.pointers %[[VAL_3]], %[[VAL_5]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_14:.*]] = sparse_tensor.indices %[[VAL_3]], %[[VAL_5]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_15:.*]] = sparse_tensor.values %[[VAL_3]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf64>
// CHECK: %[[VAL_16:.*]] = tensor.dim %[[VAL_4]], %[[VAL_5]] : tensor<?xf64>
// CHECK: %[[VAL_17:.*]] = bufferization.to_memref %[[VAL_4]] : memref<?xf64>
// CHECK: %[[VAL_18:.*]] = memref.alloc(%[[VAL_16]]) : memref<?xf64>
// CHECK: memref.copy %[[VAL_17]], %[[VAL_18]] : memref<?xf64> to memref<?xf64>
// CHECK: %[[VAL_19:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_5]]] : memref<?xindex>
// CHECK: %[[VAL_20:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_7]]] : memref<?xindex>
// CHECK: %[[VAL_21:.*]] = memref.load %[[VAL_13]]{{\[}}%[[VAL_5]]] : memref<?xindex>
// CHECK: %[[VAL_22:.*]] = memref.load %[[VAL_13]]{{\[}}%[[VAL_7]]] : memref<?xindex>
// CHECK: %[[VAL_23:.*]]:3 = scf.while (%[[VAL_24:.*]] = %[[VAL_19]], %[[VAL_25:.*]] = %[[VAL_21]], %[[VAL_26:.*]] = %[[VAL_5]]) : (index, index, index) -> (index, index, index) {
// CHECK: %[[VAL_27:.*]] = arith.cmpi ult, %[[VAL_24]], %[[VAL_20]] : index
// CHECK: %[[VAL_28:.*]] = arith.cmpi ult, %[[VAL_25]], %[[VAL_22]] : index
// CHECK: %[[VAL_29:.*]] = arith.andi %[[VAL_27]], %[[VAL_28]] : i1
// CHECK: scf.condition(%[[VAL_29]]) %[[VAL_24]], %[[VAL_25]], %[[VAL_26]] : index, index, index
// CHECK: } do {
// CHECK: ^bb0(%[[VAL_30:.*]]: index, %[[VAL_31:.*]]: index, %[[VAL_32:.*]]: index):
// CHECK: %[[VAL_33:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_30]]] : memref<?xindex>
// CHECK: %[[VAL_34:.*]] = memref.load %[[VAL_14]]{{\[}}%[[VAL_31]]] : memref<?xindex>
// CHECK: %[[VAL_35:.*]] = arith.cmpi eq, %[[VAL_33]], %[[VAL_32]] : index
// CHECK: %[[VAL_36:.*]] = arith.cmpi eq, %[[VAL_34]], %[[VAL_32]] : index
// CHECK: %[[VAL_37:.*]] = arith.andi %[[VAL_35]], %[[VAL_36]] : i1
// CHECK: scf.if %[[VAL_37]] {
// CHECK: %[[VAL_38:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_32]]] : memref<?xf64>
// CHECK: %[[VAL_39:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_30]]] : memref<?xf64>
// CHECK: %[[VAL_40:.*]] = arith.addf %[[VAL_38]], %[[VAL_39]] : f64
// CHECK: %[[VAL_41:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_32]]] : memref<?xf64>
// CHECK: %[[VAL_42:.*]] = memref.load %[[VAL_15]]{{\[}}%[[VAL_31]]] : memref<?xf64>
// CHECK: %[[VAL_43:.*]] = arith.addf %[[VAL_41]], %[[VAL_42]] : f64
// CHECK: %[[VAL_44:.*]] = arith.addf %[[VAL_40]], %[[VAL_43]] : f64
// CHECK: memref.store %[[VAL_44]], %[[VAL_18]]{{\[}}%[[VAL_32]]] : memref<?xf64>
// CHECK: } else {
// CHECK: %[[VAL_45:.*]] = arith.cmpi eq, %[[VAL_33]], %[[VAL_32]] : index
// CHECK: scf.if %[[VAL_45]] {
// CHECK: %[[VAL_46:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_32]]] : memref<?xf64>
// CHECK: %[[VAL_47:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_30]]] : memref<?xf64>
// CHECK: %[[VAL_48:.*]] = arith.addf %[[VAL_46]], %[[VAL_47]] : f64
// CHECK: %[[VAL_49:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_32]]] : memref<?xf64>
// CHECK: %[[VAL_50:.*]] = arith.addf %[[VAL_48]], %[[VAL_49]] : f64
// CHECK: memref.store %[[VAL_50]], %[[VAL_18]]{{\[}}%[[VAL_32]]] : memref<?xf64>
// CHECK: } else {
// CHECK: %[[VAL_51:.*]] = arith.cmpi eq, %[[VAL_34]], %[[VAL_32]] : index
// CHECK: scf.if %[[VAL_51]] {
// CHECK: %[[VAL_52:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_32]]] : memref<?xf64>
// CHECK: %[[VAL_53:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_32]]] : memref<?xf64>
// CHECK: %[[VAL_54:.*]] = memref.load %[[VAL_15]]{{\[}}%[[VAL_31]]] : memref<?xf64>
// CHECK: %[[VAL_55:.*]] = arith.addf %[[VAL_53]], %[[VAL_54]] : f64
// CHECK: %[[VAL_56:.*]] = arith.addf %[[VAL_52]], %[[VAL_55]] : f64
// CHECK: memref.store %[[VAL_56]], %[[VAL_18]]{{\[}}%[[VAL_32]]] : memref<?xf64>
// CHECK: } else {
// CHECK: scf.if %[[VAL_6]] {
// CHECK: %[[VAL_57:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_32]]] : memref<?xf64>
// CHECK: %[[VAL_58:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_32]]] : memref<?xf64>
// CHECK: %[[VAL_59:.*]] = arith.addf %[[VAL_57]], %[[VAL_58]] : f64
// CHECK: memref.store %[[VAL_59]], %[[VAL_18]]{{\[}}%[[VAL_32]]] : memref<?xf64>
// CHECK: } else {
// CHECK: }
// CHECK: }
// CHECK: }
// CHECK: }
// CHECK: %[[VAL_60:.*]] = arith.cmpi eq, %[[VAL_33]], %[[VAL_32]] : index
// CHECK: %[[VAL_61:.*]] = arith.addi %[[VAL_30]], %[[VAL_7]] : index
// CHECK: %[[VAL_62:.*]] = select %[[VAL_60]], %[[VAL_61]], %[[VAL_30]] : index
// CHECK: %[[VAL_63:.*]] = arith.cmpi eq, %[[VAL_34]], %[[VAL_32]] : index
// CHECK: %[[VAL_64:.*]] = arith.addi %[[VAL_31]], %[[VAL_7]] : index
// CHECK: %[[VAL_65:.*]] = select %[[VAL_63]], %[[VAL_64]], %[[VAL_31]] : index
// CHECK: %[[VAL_66:.*]] = arith.addi %[[VAL_32]], %[[VAL_7]] : index
// CHECK: scf.yield %[[VAL_62]], %[[VAL_65]], %[[VAL_66]] : index, index, index
// CHECK: }
// CHECK: %[[VAL_67:.*]]:2 = scf.while (%[[VAL_68:.*]] = %[[VAL_69:.*]]#0, %[[VAL_70:.*]] = %[[VAL_69]]#2) : (index, index) -> (index, index) {
// CHECK: %[[VAL_71:.*]] = arith.cmpi ult, %[[VAL_68]], %[[VAL_20]] : index
// CHECK: scf.condition(%[[VAL_71]]) %[[VAL_68]], %[[VAL_70]] : index, index
// CHECK: } do {
// CHECK: ^bb0(%[[VAL_72:.*]]: index, %[[VAL_73:.*]]: index):
// CHECK: %[[VAL_74:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_72]]] : memref<?xindex>
// CHECK: %[[VAL_75:.*]] = arith.cmpi eq, %[[VAL_74]], %[[VAL_73]] : index
// CHECK: scf.if %[[VAL_75]] {
// CHECK: %[[VAL_76:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_73]]] : memref<?xf64>
// CHECK: %[[VAL_77:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_72]]] : memref<?xf64>
// CHECK: %[[VAL_78:.*]] = arith.addf %[[VAL_76]], %[[VAL_77]] : f64
// CHECK: %[[VAL_79:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_73]]] : memref<?xf64>
// CHECK: %[[VAL_80:.*]] = arith.addf %[[VAL_78]], %[[VAL_79]] : f64
// CHECK: memref.store %[[VAL_80]], %[[VAL_18]]{{\[}}%[[VAL_73]]] : memref<?xf64>
// CHECK: } else {
// CHECK: scf.if %[[VAL_6]] {
// CHECK: %[[VAL_81:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_73]]] : memref<?xf64>
// CHECK: %[[VAL_82:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_73]]] : memref<?xf64>
// CHECK: %[[VAL_83:.*]] = arith.addf %[[VAL_81]], %[[VAL_82]] : f64
// CHECK: memref.store %[[VAL_83]], %[[VAL_18]]{{\[}}%[[VAL_73]]] : memref<?xf64>
// CHECK: } else {
// CHECK: }
// CHECK: }
// CHECK: %[[VAL_84:.*]] = arith.cmpi eq, %[[VAL_74]], %[[VAL_73]] : index
// CHECK: %[[VAL_85:.*]] = arith.addi %[[VAL_72]], %[[VAL_7]] : index
// CHECK: %[[VAL_86:.*]] = select %[[VAL_84]], %[[VAL_85]], %[[VAL_72]] : index
// CHECK: %[[VAL_87:.*]] = arith.addi %[[VAL_73]], %[[VAL_7]] : index
// CHECK: scf.yield %[[VAL_86]], %[[VAL_87]] : index, index
// CHECK: }
// CHECK: %[[VAL_88:.*]]:2 = scf.while (%[[VAL_89:.*]] = %[[VAL_90:.*]]#1, %[[VAL_91:.*]] = %[[VAL_92:.*]]#1) : (index, index) -> (index, index) {
// CHECK: %[[VAL_93:.*]] = arith.cmpi ult, %[[VAL_89]], %[[VAL_22]] : index
// CHECK: scf.condition(%[[VAL_93]]) %[[VAL_89]], %[[VAL_91]] : index, index
// CHECK: } do {
// CHECK: ^bb0(%[[VAL_94:.*]]: index, %[[VAL_95:.*]]: index):
// CHECK: %[[VAL_96:.*]] = memref.load %[[VAL_14]]{{\[}}%[[VAL_94]]] : memref<?xindex>
// CHECK: %[[VAL_97:.*]] = arith.cmpi eq, %[[VAL_96]], %[[VAL_95]] : index
// CHECK: scf.if %[[VAL_97]] {
// CHECK: %[[VAL_98:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_95]]] : memref<?xf64>
// CHECK: %[[VAL_99:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_95]]] : memref<?xf64>
// CHECK: %[[VAL_100:.*]] = memref.load %[[VAL_15]]{{\[}}%[[VAL_94]]] : memref<?xf64>
// CHECK: %[[VAL_101:.*]] = arith.addf %[[VAL_99]], %[[VAL_100]] : f64
// CHECK: %[[VAL_102:.*]] = arith.addf %[[VAL_98]], %[[VAL_101]] : f64
// CHECK: memref.store %[[VAL_102]], %[[VAL_18]]{{\[}}%[[VAL_95]]] : memref<?xf64>
// CHECK: } else {
// CHECK: scf.if %[[VAL_6]] {
// CHECK: %[[VAL_103:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_95]]] : memref<?xf64>
// CHECK: %[[VAL_104:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_95]]] : memref<?xf64>
// CHECK: %[[VAL_105:.*]] = arith.addf %[[VAL_103]], %[[VAL_104]] : f64
// CHECK: memref.store %[[VAL_105]], %[[VAL_18]]{{\[}}%[[VAL_95]]] : memref<?xf64>
// CHECK: } else {
// CHECK: }
// CHECK: }
// CHECK: %[[VAL_106:.*]] = arith.cmpi eq, %[[VAL_96]], %[[VAL_95]] : index
// CHECK: %[[VAL_107:.*]] = arith.addi %[[VAL_94]], %[[VAL_7]] : index
// CHECK: %[[VAL_108:.*]] = select %[[VAL_106]], %[[VAL_107]], %[[VAL_94]] : index
// CHECK: %[[VAL_109:.*]] = arith.addi %[[VAL_95]], %[[VAL_7]] : index
// CHECK: scf.yield %[[VAL_108]], %[[VAL_109]] : index, index
// CHECK: }
// CHECK: scf.for %[[VAL_110:.*]] = %[[VAL_111:.*]]#1 to %[[VAL_16]] step %[[VAL_7]] {
// CHECK: %[[VAL_112:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_110]]] : memref<?xf64>
// CHECK: %[[VAL_113:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_110]]] : memref<?xf64>
// CHECK: %[[VAL_114:.*]] = arith.addf %[[VAL_112]], %[[VAL_113]] : f64
// CHECK: memref.store %[[VAL_114]], %[[VAL_18]]{{\[}}%[[VAL_110]]] : memref<?xf64>
// CHECK: }
// CHECK: %[[VAL_115:.*]] = bufferization.to_tensor %[[VAL_18]] : memref<?xf64>
// CHECK: return %[[VAL_115]] : tensor<?xf64>
// CHECK: }
func @four_tensors_op(%arga: tensor<?xf64>,
%argb: tensor<?xf64, #SV>,
%argc: tensor<?xf64>,
%argd: tensor<?xf64, #SV>,
%argx: tensor<?xf64>) -> tensor<?xf64> {
%r = linalg.generic #trait_four_tensors
ins(%arga, %argb, %argc, %argd: tensor<?xf64>, tensor<?xf64, #SV>, tensor<?xf64>, tensor<?xf64, #SV>)
outs(%argx: tensor<?xf64>) {
^bb(%a: f64, %b: f64, %c: f64, %d: f64, %x: f64):
%0 = arith.addf %a, %b : f64
%1 = arith.addf %c, %d : f64
%2 = arith.addf %0, %1 : f64
linalg.yield %2 : f64
} -> tensor<?xf64>
return %r : tensor<?xf64>
}
#trait_red3s = {
indexing_maps = [
affine_map<(i) -> (i)>,
affine_map<(i) -> (i)>,
affine_map<(i) -> (i)>,
affine_map<(i) -> ()>
],
iterator_types = ["reduction"],
doc = "x += a(i) + b(i) + c(i)"
}
// CHECK-LABEL: func @red3s(
// CHECK-SAME: %[[VAL_0:.*0]]: tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_1:.*1]]: tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_2:.*2]]: tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>>,
// CHECK-SAME: %[[VAL_3:.*3]]: tensor<f64>) -> tensor<f64> {
// CHECK-DAG: %[[VAL_4:.*]] = arith.constant 0 : index
// CHECK-DAG: %[[VAL_5:.*]] = arith.constant 1 : index
// CHECK: %[[VAL_6:.*]] = sparse_tensor.pointers %[[VAL_0]], %[[VAL_4]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_7:.*]] = sparse_tensor.indices %[[VAL_0]], %[[VAL_4]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_8:.*]] = sparse_tensor.values %[[VAL_0]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf64>
// CHECK: %[[VAL_9:.*]] = sparse_tensor.pointers %[[VAL_1]], %[[VAL_4]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_10:.*]] = sparse_tensor.indices %[[VAL_1]], %[[VAL_4]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_11:.*]] = sparse_tensor.values %[[VAL_1]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf64>
// CHECK: %[[VAL_12:.*]] = sparse_tensor.pointers %[[VAL_2]], %[[VAL_4]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_13:.*]] = sparse_tensor.indices %[[VAL_2]], %[[VAL_4]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xindex>
// CHECK: %[[VAL_14:.*]] = sparse_tensor.values %[[VAL_2]] : tensor<?xf64, #sparse_tensor.encoding<{ dimLevelType = [ "compressed" ], pointerBitWidth = 0, indexBitWidth = 0 }>> to memref<?xf64>
// CHECK: %[[VAL_15:.*]] = bufferization.to_memref %[[VAL_3]] : memref<f64>
// CHECK: %[[VAL_16:.*]] = memref.alloc() : memref<f64>
// CHECK: memref.copy %[[VAL_15]], %[[VAL_16]] : memref<f64> to memref<f64>
// CHECK: %[[VAL_17:.*]] = memref.load %[[VAL_16]][] : memref<f64>
// CHECK: %[[VAL_18:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_4]]] : memref<?xindex>
// CHECK: %[[VAL_19:.*]] = memref.load %[[VAL_6]]{{\[}}%[[VAL_5]]] : memref<?xindex>
// CHECK: %[[VAL_20:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_4]]] : memref<?xindex>
// CHECK: %[[VAL_21:.*]] = memref.load %[[VAL_9]]{{\[}}%[[VAL_5]]] : memref<?xindex>
// CHECK: %[[VAL_22:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_4]]] : memref<?xindex>
// CHECK: %[[VAL_23:.*]] = memref.load %[[VAL_12]]{{\[}}%[[VAL_5]]] : memref<?xindex>
// CHECK: %[[VAL_24:.*]]:4 = scf.while (%[[VAL_25:.*]] = %[[VAL_18]], %[[VAL_26:.*]] = %[[VAL_20]], %[[VAL_27:.*]] = %[[VAL_22]], %[[VAL_28:.*]] = %[[VAL_17]]) : (index, index, index, f64) -> (index, index, index, f64) {
// CHECK: %[[VAL_29:.*]] = arith.cmpi ult, %[[VAL_25]], %[[VAL_19]] : index
// CHECK: %[[VAL_30:.*]] = arith.cmpi ult, %[[VAL_26]], %[[VAL_21]] : index
// CHECK: %[[VAL_31:.*]] = arith.andi %[[VAL_29]], %[[VAL_30]] : i1
// CHECK: %[[VAL_32:.*]] = arith.cmpi ult, %[[VAL_27]], %[[VAL_23]] : index
// CHECK: %[[VAL_33:.*]] = arith.andi %[[VAL_31]], %[[VAL_32]] : i1
// CHECK: scf.condition(%[[VAL_33]]) %[[VAL_25]], %[[VAL_26]], %[[VAL_27]], %[[VAL_28]] : index, index, index, f64
// CHECK: } do {
// CHECK: ^bb0(%[[VAL_34:.*]]: index, %[[VAL_35:.*]]: index, %[[VAL_36:.*]]: index, %[[VAL_37:.*]]: f64):
// CHECK: %[[VAL_38:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_34]]] : memref<?xindex>
// CHECK: %[[VAL_39:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_35]]] : memref<?xindex>
// CHECK: %[[VAL_40:.*]] = arith.cmpi ult, %[[VAL_39]], %[[VAL_38]] : index
// CHECK: %[[VAL_41:.*]] = select %[[VAL_40]], %[[VAL_39]], %[[VAL_38]] : index
// CHECK: %[[VAL_42:.*]] = memref.load %[[VAL_13]]{{\[}}%[[VAL_36]]] : memref<?xindex>
// CHECK: %[[VAL_43:.*]] = arith.cmpi ult, %[[VAL_42]], %[[VAL_41]] : index
// CHECK: %[[VAL_44:.*]] = select %[[VAL_43]], %[[VAL_42]], %[[VAL_41]] : index
// CHECK: %[[VAL_45:.*]] = arith.cmpi eq, %[[VAL_38]], %[[VAL_44]] : index
// CHECK: %[[VAL_46:.*]] = arith.cmpi eq, %[[VAL_39]], %[[VAL_44]] : index
// CHECK: %[[VAL_47:.*]] = arith.andi %[[VAL_45]], %[[VAL_46]] : i1
// CHECK: %[[VAL_48:.*]] = arith.cmpi eq, %[[VAL_42]], %[[VAL_44]] : index
// CHECK: %[[VAL_49:.*]] = arith.andi %[[VAL_47]], %[[VAL_48]] : i1
// CHECK: %[[VAL_50:.*]] = scf.if %[[VAL_49]] -> (f64) {
// CHECK: %[[VAL_51:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_34]]] : memref<?xf64>
// CHECK: %[[VAL_52:.*]] = arith.addf %[[VAL_37]], %[[VAL_51]] : f64
// CHECK: %[[VAL_53:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_35]]] : memref<?xf64>
// CHECK: %[[VAL_54:.*]] = arith.addf %[[VAL_52]], %[[VAL_53]] : f64
// CHECK: %[[VAL_55:.*]] = memref.load %[[VAL_14]]{{\[}}%[[VAL_36]]] : memref<?xf64>
// CHECK: %[[VAL_56:.*]] = arith.addf %[[VAL_54]], %[[VAL_55]] : f64
// CHECK: scf.yield %[[VAL_56]] : f64
// CHECK: } else {
// CHECK: %[[VAL_57:.*]] = arith.cmpi eq, %[[VAL_39]], %[[VAL_44]] : index
// CHECK: %[[VAL_58:.*]] = arith.cmpi eq, %[[VAL_42]], %[[VAL_44]] : index
// CHECK: %[[VAL_59:.*]] = arith.andi %[[VAL_57]], %[[VAL_58]] : i1
// CHECK: %[[VAL_60:.*]] = scf.if %[[VAL_59]] -> (f64) {
// CHECK: %[[VAL_61:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_35]]] : memref<?xf64>
// CHECK: %[[VAL_62:.*]] = arith.addf %[[VAL_37]], %[[VAL_61]] : f64
// CHECK: %[[VAL_63:.*]] = memref.load %[[VAL_14]]{{\[}}%[[VAL_36]]] : memref<?xf64>
// CHECK: %[[VAL_64:.*]] = arith.addf %[[VAL_62]], %[[VAL_63]] : f64
// CHECK: scf.yield %[[VAL_64]] : f64
// CHECK: } else {
// CHECK: %[[VAL_65:.*]] = arith.cmpi eq, %[[VAL_38]], %[[VAL_44]] : index
// CHECK: %[[VAL_66:.*]] = arith.cmpi eq, %[[VAL_42]], %[[VAL_44]] : index
// CHECK: %[[VAL_67:.*]] = arith.andi %[[VAL_65]], %[[VAL_66]] : i1
// CHECK: %[[VAL_68:.*]] = scf.if %[[VAL_67]] -> (f64) {
// CHECK: %[[VAL_69:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_34]]] : memref<?xf64>
// CHECK: %[[VAL_70:.*]] = arith.addf %[[VAL_37]], %[[VAL_69]] : f64
// CHECK: %[[VAL_71:.*]] = memref.load %[[VAL_14]]{{\[}}%[[VAL_36]]] : memref<?xf64>
// CHECK: %[[VAL_72:.*]] = arith.addf %[[VAL_70]], %[[VAL_71]] : f64
// CHECK: scf.yield %[[VAL_72]] : f64
// CHECK: } else {
// CHECK: %[[VAL_73:.*]] = arith.cmpi eq, %[[VAL_42]], %[[VAL_44]] : index
// CHECK: %[[VAL_74:.*]] = scf.if %[[VAL_73]] -> (f64) {
// CHECK: %[[VAL_75:.*]] = memref.load %[[VAL_14]]{{\[}}%[[VAL_36]]] : memref<?xf64>
// CHECK: %[[VAL_76:.*]] = arith.addf %[[VAL_37]], %[[VAL_75]] : f64
// CHECK: scf.yield %[[VAL_76]] : f64
// CHECK: } else {
// CHECK: %[[VAL_77:.*]] = arith.cmpi eq, %[[VAL_38]], %[[VAL_44]] : index
// CHECK: %[[VAL_78:.*]] = arith.cmpi eq, %[[VAL_39]], %[[VAL_44]] : index
// CHECK: %[[VAL_79:.*]] = arith.andi %[[VAL_77]], %[[VAL_78]] : i1
// CHECK: %[[VAL_80:.*]] = scf.if %[[VAL_79]] -> (f64) {
// CHECK: %[[VAL_81:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_34]]] : memref<?xf64>
// CHECK: %[[VAL_82:.*]] = arith.addf %[[VAL_37]], %[[VAL_81]] : f64
// CHECK: %[[VAL_83:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_35]]] : memref<?xf64>
// CHECK: %[[VAL_84:.*]] = arith.addf %[[VAL_82]], %[[VAL_83]] : f64
// CHECK: scf.yield %[[VAL_84]] : f64
// CHECK: } else {
// CHECK: %[[VAL_85:.*]] = arith.cmpi eq, %[[VAL_39]], %[[VAL_44]] : index
// CHECK: %[[VAL_86:.*]] = scf.if %[[VAL_85]] -> (f64) {
// CHECK: %[[VAL_87:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_35]]] : memref<?xf64>
// CHECK: %[[VAL_88:.*]] = arith.addf %[[VAL_37]], %[[VAL_87]] : f64
// CHECK: scf.yield %[[VAL_88]] : f64
// CHECK: } else {
// CHECK: %[[VAL_89:.*]] = arith.cmpi eq, %[[VAL_38]], %[[VAL_44]] : index
// CHECK: %[[VAL_90:.*]] = scf.if %[[VAL_89]] -> (f64) {
// CHECK: %[[VAL_91:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_34]]] : memref<?xf64>
// CHECK: %[[VAL_92:.*]] = arith.addf %[[VAL_37]], %[[VAL_91]] : f64
// CHECK: scf.yield %[[VAL_92]] : f64
// CHECK: } else {
// CHECK: scf.yield %[[VAL_37]] : f64
// CHECK: }
// CHECK: scf.yield %[[VAL_93:.*]] : f64
// CHECK: }
// CHECK: scf.yield %[[VAL_94:.*]] : f64
// CHECK: }
// CHECK: scf.yield %[[VAL_95:.*]] : f64
// CHECK: }
// CHECK: scf.yield %[[VAL_96:.*]] : f64
// CHECK: }
// CHECK: scf.yield %[[VAL_97:.*]] : f64
// CHECK: }
// CHECK: scf.yield %[[VAL_98:.*]] : f64
// CHECK: }
// CHECK: %[[VAL_99:.*]] = arith.cmpi eq, %[[VAL_38]], %[[VAL_44]] : index
// CHECK: %[[VAL_100:.*]] = arith.addi %[[VAL_34]], %[[VAL_5]] : index
// CHECK: %[[VAL_101:.*]] = select %[[VAL_99]], %[[VAL_100]], %[[VAL_34]] : index
// CHECK: %[[VAL_102:.*]] = arith.cmpi eq, %[[VAL_39]], %[[VAL_44]] : index
// CHECK: %[[VAL_103:.*]] = arith.addi %[[VAL_35]], %[[VAL_5]] : index
// CHECK: %[[VAL_104:.*]] = select %[[VAL_102]], %[[VAL_103]], %[[VAL_35]] : index
// CHECK: %[[VAL_105:.*]] = arith.cmpi eq, %[[VAL_42]], %[[VAL_44]] : index
// CHECK: %[[VAL_106:.*]] = arith.addi %[[VAL_36]], %[[VAL_5]] : index
// CHECK: %[[VAL_107:.*]] = select %[[VAL_105]], %[[VAL_106]], %[[VAL_36]] : index
// CHECK: scf.yield %[[VAL_101]], %[[VAL_104]], %[[VAL_107]], %[[VAL_108:.*]] : index, index, index, f64
// CHECK: }
// CHECK: %[[VAL_109:.*]]:3 = scf.while (%[[VAL_110:.*]] = %[[VAL_111:.*]]#1, %[[VAL_112:.*]] = %[[VAL_111]]#2, %[[VAL_113:.*]] = %[[VAL_111]]#3) : (index, index, f64) -> (index, index, f64) {
// CHECK: %[[VAL_114:.*]] = arith.cmpi ult, %[[VAL_110]], %[[VAL_21]] : index
// CHECK: %[[VAL_115:.*]] = arith.cmpi ult, %[[VAL_112]], %[[VAL_23]] : index
// CHECK: %[[VAL_116:.*]] = arith.andi %[[VAL_114]], %[[VAL_115]] : i1
// CHECK: scf.condition(%[[VAL_116]]) %[[VAL_110]], %[[VAL_112]], %[[VAL_113]] : index, index, f64
// CHECK: } do {
// CHECK: ^bb0(%[[VAL_117:.*]]: index, %[[VAL_118:.*]]: index, %[[VAL_119:.*]]: f64):
// CHECK: %[[VAL_120:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_117]]] : memref<?xindex>
// CHECK: %[[VAL_121:.*]] = memref.load %[[VAL_13]]{{\[}}%[[VAL_118]]] : memref<?xindex>
// CHECK: %[[VAL_122:.*]] = arith.cmpi ult, %[[VAL_121]], %[[VAL_120]] : index
// CHECK: %[[VAL_123:.*]] = select %[[VAL_122]], %[[VAL_121]], %[[VAL_120]] : index
// CHECK: %[[VAL_124:.*]] = arith.cmpi eq, %[[VAL_120]], %[[VAL_123]] : index
// CHECK: %[[VAL_125:.*]] = arith.cmpi eq, %[[VAL_121]], %[[VAL_123]] : index
// CHECK: %[[VAL_126:.*]] = arith.andi %[[VAL_124]], %[[VAL_125]] : i1
// CHECK: %[[VAL_127:.*]] = scf.if %[[VAL_126]] -> (f64) {
// CHECK: %[[VAL_128:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_117]]] : memref<?xf64>
// CHECK: %[[VAL_129:.*]] = arith.addf %[[VAL_119]], %[[VAL_128]] : f64
// CHECK: %[[VAL_130:.*]] = memref.load %[[VAL_14]]{{\[}}%[[VAL_118]]] : memref<?xf64>
// CHECK: %[[VAL_131:.*]] = arith.addf %[[VAL_129]], %[[VAL_130]] : f64
// CHECK: scf.yield %[[VAL_131]] : f64
// CHECK: } else {
// CHECK: %[[VAL_132:.*]] = arith.cmpi eq, %[[VAL_121]], %[[VAL_123]] : index
// CHECK: %[[VAL_133:.*]] = scf.if %[[VAL_132]] -> (f64) {
// CHECK: %[[VAL_134:.*]] = memref.load %[[VAL_14]]{{\[}}%[[VAL_118]]] : memref<?xf64>
// CHECK: %[[VAL_135:.*]] = arith.addf %[[VAL_119]], %[[VAL_134]] : f64
// CHECK: scf.yield %[[VAL_135]] : f64
// CHECK: } else {
// CHECK: %[[VAL_136:.*]] = arith.cmpi eq, %[[VAL_120]], %[[VAL_123]] : index
// CHECK: %[[VAL_137:.*]] = scf.if %[[VAL_136]] -> (f64) {
// CHECK: %[[VAL_138:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_117]]] : memref<?xf64>
// CHECK: %[[VAL_139:.*]] = arith.addf %[[VAL_119]], %[[VAL_138]] : f64
// CHECK: scf.yield %[[VAL_139]] : f64
// CHECK: } else {
// CHECK: scf.yield %[[VAL_119]] : f64
// CHECK: }
// CHECK: scf.yield %[[VAL_140:.*]] : f64
// CHECK: }
// CHECK: scf.yield %[[VAL_141:.*]] : f64
// CHECK: }
// CHECK: %[[VAL_142:.*]] = arith.cmpi eq, %[[VAL_120]], %[[VAL_123]] : index
// CHECK: %[[VAL_143:.*]] = arith.addi %[[VAL_117]], %[[VAL_5]] : index
// CHECK: %[[VAL_144:.*]] = select %[[VAL_142]], %[[VAL_143]], %[[VAL_117]] : index
// CHECK: %[[VAL_145:.*]] = arith.cmpi eq, %[[VAL_121]], %[[VAL_123]] : index
// CHECK: %[[VAL_146:.*]] = arith.addi %[[VAL_118]], %[[VAL_5]] : index
// CHECK: %[[VAL_147:.*]] = select %[[VAL_145]], %[[VAL_146]], %[[VAL_118]] : index
// CHECK: scf.yield %[[VAL_144]], %[[VAL_147]], %[[VAL_148:.*]] : index, index, f64
// CHECK: }
// CHECK: %[[VAL_149:.*]]:3 = scf.while (%[[VAL_150:.*]] = %[[VAL_151:.*]]#0, %[[VAL_152:.*]] = %[[VAL_153:.*]]#1, %[[VAL_154:.*]] = %[[VAL_153]]#2) : (index, index, f64) -> (index, index, f64) {
// CHECK: %[[VAL_155:.*]] = arith.cmpi ult, %[[VAL_150]], %[[VAL_19]] : index
// CHECK: %[[VAL_156:.*]] = arith.cmpi ult, %[[VAL_152]], %[[VAL_23]] : index
// CHECK: %[[VAL_157:.*]] = arith.andi %[[VAL_155]], %[[VAL_156]] : i1
// CHECK: scf.condition(%[[VAL_157]]) %[[VAL_150]], %[[VAL_152]], %[[VAL_154]] : index, index, f64
// CHECK: } do {
// CHECK: ^bb0(%[[VAL_158:.*]]: index, %[[VAL_159:.*]]: index, %[[VAL_160:.*]]: f64):
// CHECK: %[[VAL_161:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_158]]] : memref<?xindex>
// CHECK: %[[VAL_162:.*]] = memref.load %[[VAL_13]]{{\[}}%[[VAL_159]]] : memref<?xindex>
// CHECK: %[[VAL_163:.*]] = arith.cmpi ult, %[[VAL_162]], %[[VAL_161]] : index
// CHECK: %[[VAL_164:.*]] = select %[[VAL_163]], %[[VAL_162]], %[[VAL_161]] : index
// CHECK: %[[VAL_165:.*]] = arith.cmpi eq, %[[VAL_161]], %[[VAL_164]] : index
// CHECK: %[[VAL_166:.*]] = arith.cmpi eq, %[[VAL_162]], %[[VAL_164]] : index
// CHECK: %[[VAL_167:.*]] = arith.andi %[[VAL_165]], %[[VAL_166]] : i1
// CHECK: %[[VAL_168:.*]] = scf.if %[[VAL_167]] -> (f64) {
// CHECK: %[[VAL_169:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_158]]] : memref<?xf64>
// CHECK: %[[VAL_170:.*]] = arith.addf %[[VAL_160]], %[[VAL_169]] : f64
// CHECK: %[[VAL_171:.*]] = memref.load %[[VAL_14]]{{\[}}%[[VAL_159]]] : memref<?xf64>
// CHECK: %[[VAL_172:.*]] = arith.addf %[[VAL_170]], %[[VAL_171]] : f64
// CHECK: scf.yield %[[VAL_172]] : f64
// CHECK: } else {
// CHECK: %[[VAL_173:.*]] = arith.cmpi eq, %[[VAL_162]], %[[VAL_164]] : index
// CHECK: %[[VAL_174:.*]] = scf.if %[[VAL_173]] -> (f64) {
// CHECK: %[[VAL_175:.*]] = memref.load %[[VAL_14]]{{\[}}%[[VAL_159]]] : memref<?xf64>
// CHECK: %[[VAL_176:.*]] = arith.addf %[[VAL_160]], %[[VAL_175]] : f64
// CHECK: scf.yield %[[VAL_176]] : f64
// CHECK: } else {
// CHECK: %[[VAL_177:.*]] = arith.cmpi eq, %[[VAL_161]], %[[VAL_164]] : index
// CHECK: %[[VAL_178:.*]] = scf.if %[[VAL_177]] -> (f64) {
// CHECK: %[[VAL_179:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_158]]] : memref<?xf64>
// CHECK: %[[VAL_180:.*]] = arith.addf %[[VAL_160]], %[[VAL_179]] : f64
// CHECK: scf.yield %[[VAL_180]] : f64
// CHECK: } else {
// CHECK: scf.yield %[[VAL_160]] : f64
// CHECK: }
// CHECK: scf.yield %[[VAL_181:.*]] : f64
// CHECK: }
// CHECK: scf.yield %[[VAL_182:.*]] : f64
// CHECK: }
// CHECK: %[[VAL_183:.*]] = arith.cmpi eq, %[[VAL_161]], %[[VAL_164]] : index
// CHECK: %[[VAL_184:.*]] = arith.addi %[[VAL_158]], %[[VAL_5]] : index
// CHECK: %[[VAL_185:.*]] = select %[[VAL_183]], %[[VAL_184]], %[[VAL_158]] : index
// CHECK: %[[VAL_186:.*]] = arith.cmpi eq, %[[VAL_162]], %[[VAL_164]] : index
// CHECK: %[[VAL_187:.*]] = arith.addi %[[VAL_159]], %[[VAL_5]] : index
// CHECK: %[[VAL_188:.*]] = select %[[VAL_186]], %[[VAL_187]], %[[VAL_159]] : index
// CHECK: scf.yield %[[VAL_185]], %[[VAL_188]], %[[VAL_189:.*]] : index, index, f64
// CHECK: }
// CHECK: %[[VAL_190:.*]] = scf.for %[[VAL_191:.*]] = %[[VAL_192:.*]]#1 to %[[VAL_23]] step %[[VAL_5]] iter_args(%[[VAL_193:.*]] = %[[VAL_192]]#2) -> (f64) {
// CHECK: %[[VAL_194:.*]] = memref.load %[[VAL_14]]{{\[}}%[[VAL_191]]] : memref<?xf64>
// CHECK: %[[VAL_195:.*]] = arith.addf %[[VAL_193]], %[[VAL_194]] : f64
// CHECK: scf.yield %[[VAL_195]] : f64
// CHECK: }
// CHECK: %[[VAL_196:.*]]:3 = scf.while (%[[VAL_197:.*]] = %[[VAL_198:.*]]#0, %[[VAL_199:.*]] = %[[VAL_200:.*]]#0, %[[VAL_201:.*]] = %[[VAL_202:.*]]) : (index, index, f64) -> (index, index, f64) {
// CHECK: %[[VAL_203:.*]] = arith.cmpi ult, %[[VAL_197]], %[[VAL_19]] : index
// CHECK: %[[VAL_204:.*]] = arith.cmpi ult, %[[VAL_199]], %[[VAL_21]] : index
// CHECK: %[[VAL_205:.*]] = arith.andi %[[VAL_203]], %[[VAL_204]] : i1
// CHECK: scf.condition(%[[VAL_205]]) %[[VAL_197]], %[[VAL_199]], %[[VAL_201]] : index, index, f64
// CHECK: } do {
// CHECK: ^bb0(%[[VAL_206:.*]]: index, %[[VAL_207:.*]]: index, %[[VAL_208:.*]]: f64):
// CHECK: %[[VAL_209:.*]] = memref.load %[[VAL_7]]{{\[}}%[[VAL_206]]] : memref<?xindex>
// CHECK: %[[VAL_210:.*]] = memref.load %[[VAL_10]]{{\[}}%[[VAL_207]]] : memref<?xindex>
// CHECK: %[[VAL_211:.*]] = arith.cmpi ult, %[[VAL_210]], %[[VAL_209]] : index
// CHECK: %[[VAL_212:.*]] = select %[[VAL_211]], %[[VAL_210]], %[[VAL_209]] : index
// CHECK: %[[VAL_213:.*]] = arith.cmpi eq, %[[VAL_209]], %[[VAL_212]] : index
// CHECK: %[[VAL_214:.*]] = arith.cmpi eq, %[[VAL_210]], %[[VAL_212]] : index
// CHECK: %[[VAL_215:.*]] = arith.andi %[[VAL_213]], %[[VAL_214]] : i1
// CHECK: %[[VAL_216:.*]] = scf.if %[[VAL_215]] -> (f64) {
// CHECK: %[[VAL_217:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_206]]] : memref<?xf64>
// CHECK: %[[VAL_218:.*]] = arith.addf %[[VAL_208]], %[[VAL_217]] : f64
// CHECK: %[[VAL_219:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_207]]] : memref<?xf64>
// CHECK: %[[VAL_220:.*]] = arith.addf %[[VAL_218]], %[[VAL_219]] : f64
// CHECK: scf.yield %[[VAL_220]] : f64
// CHECK: } else {
// CHECK: %[[VAL_221:.*]] = arith.cmpi eq, %[[VAL_210]], %[[VAL_212]] : index
// CHECK: %[[VAL_222:.*]] = scf.if %[[VAL_221]] -> (f64) {
// CHECK: %[[VAL_223:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_207]]] : memref<?xf64>
// CHECK: %[[VAL_224:.*]] = arith.addf %[[VAL_208]], %[[VAL_223]] : f64
// CHECK: scf.yield %[[VAL_224]] : f64
// CHECK: } else {
// CHECK: %[[VAL_225:.*]] = arith.cmpi eq, %[[VAL_209]], %[[VAL_212]] : index
// CHECK: %[[VAL_226:.*]] = scf.if %[[VAL_225]] -> (f64) {
// CHECK: %[[VAL_227:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_206]]] : memref<?xf64>
// CHECK: %[[VAL_228:.*]] = arith.addf %[[VAL_208]], %[[VAL_227]] : f64
// CHECK: scf.yield %[[VAL_228]] : f64
// CHECK: } else {
// CHECK: scf.yield %[[VAL_208]] : f64
// CHECK: }
// CHECK: scf.yield %[[VAL_229:.*]] : f64
// CHECK: }
// CHECK: scf.yield %[[VAL_230:.*]] : f64
// CHECK: }
// CHECK: %[[VAL_231:.*]] = arith.cmpi eq, %[[VAL_209]], %[[VAL_212]] : index
// CHECK: %[[VAL_232:.*]] = arith.addi %[[VAL_206]], %[[VAL_5]] : index
// CHECK: %[[VAL_233:.*]] = select %[[VAL_231]], %[[VAL_232]], %[[VAL_206]] : index
// CHECK: %[[VAL_234:.*]] = arith.cmpi eq, %[[VAL_210]], %[[VAL_212]] : index
// CHECK: %[[VAL_235:.*]] = arith.addi %[[VAL_207]], %[[VAL_5]] : index
// CHECK: %[[VAL_236:.*]] = select %[[VAL_234]], %[[VAL_235]], %[[VAL_207]] : index
// CHECK: scf.yield %[[VAL_233]], %[[VAL_236]], %[[VAL_237:.*]] : index, index, f64
// CHECK: }
// CHECK: %[[VAL_238:.*]] = scf.for %[[VAL_239:.*]] = %[[VAL_240:.*]]#1 to %[[VAL_21]] step %[[VAL_5]] iter_args(%[[VAL_241:.*]] = %[[VAL_240]]#2) -> (f64) {
// CHECK: %[[VAL_242:.*]] = memref.load %[[VAL_11]]{{\[}}%[[VAL_239]]] : memref<?xf64>
// CHECK: %[[VAL_243:.*]] = arith.addf %[[VAL_241]], %[[VAL_242]] : f64
// CHECK: scf.yield %[[VAL_243]] : f64
// CHECK: }
// CHECK: %[[VAL_244:.*]] = scf.for %[[VAL_245:.*]] = %[[VAL_246:.*]]#0 to %[[VAL_19]] step %[[VAL_5]] iter_args(%[[VAL_247:.*]] = %[[VAL_248:.*]]) -> (f64) {
// CHECK: %[[VAL_249:.*]] = memref.load %[[VAL_8]]{{\[}}%[[VAL_245]]] : memref<?xf64>
// CHECK: %[[VAL_250:.*]] = arith.addf %[[VAL_247]], %[[VAL_249]] : f64
// CHECK: scf.yield %[[VAL_250]] : f64
// CHECK: }
// CHECK: memref.store %[[VAL_251:.*]], %[[VAL_16]][] : memref<f64>
// CHECK: %[[VAL_252:.*]] = bufferization.to_tensor %[[VAL_16]] : memref<f64>
// CHECK: return %[[VAL_252]] : tensor<f64>
// CHECK: }
func @red3s(%arga: tensor<?xf64, #SV>,
%argb: tensor<?xf64, #SV>,
%argc: tensor<?xf64, #SV>, %argx: tensor<f64>) ->tensor<f64>{
%0 = linalg.generic #trait_red3s
ins(%arga, %argb, %argc: tensor<?xf64, #SV>, tensor<?xf64, #SV>, tensor<?xf64, #SV>)
outs(%argx: tensor<f64>) {
^bb(%a: f64,%b: f64,%c: f64,%x: f64):
%0 = arith.addf %x, %a : f64
%1 = arith.addf %0, %b : f64
%2 = arith.addf %1, %c : f64
linalg.yield %2 : f64
} -> tensor<f64>
return %0 : tensor<f64>
}