%3 = "tosa.mul"(%arg0, %arg0) {shift = 2 : i32} : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
// CHECK: linalg.generic
+ // CHECK: divi
+ %4 = "tosa.div"(%arg0, %arg0) : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
+
+ // CHECK: linalg.generic
// CHECK: [[ZERO:%.+]] = constant 0
// CHECK: subi [[ZERO]], %arg1
- %4 = "tosa.negate"(%arg0) : (tensor<1xi32>) -> tensor<1xi32>
+ %5 = "tosa.negate"(%arg0) : (tensor<1xi32>) -> tensor<1xi32>
// CHECK: linalg.generic
// CHECK: and
- %5 = "tosa.bitwise_and"(%arg0, %arg0) : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
+ %6 = "tosa.bitwise_and"(%arg0, %arg0) : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
// CHECK: linalg.generic
// CHECK: or
- %6 = "tosa.bitwise_or"(%arg0, %arg0) : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
+ %7 = "tosa.bitwise_or"(%arg0, %arg0) : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
// CHECK: linalg.generic
// CHECK: xor
- %7 = "tosa.bitwise_xor"(%arg0, %arg0) : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
+ %8 = "tosa.bitwise_xor"(%arg0, %arg0) : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
// CHECK: linalg.generic
// CHECK: shift_left
- %8 = "tosa.logical_left_shift"(%arg0, %arg0) : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
+ %9 = "tosa.logical_left_shift"(%arg0, %arg0) : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
// CHECK: linalg.generic
// CHECK: shift_right_unsigned
- %9 = "tosa.logical_right_shift"(%arg0, %arg0) : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
+ %10 = "tosa.logical_right_shift"(%arg0, %arg0) : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
// CHECK: linalg.generic
// CHECK: shift_right_signed
- %10 = "tosa.arithmetic_right_shift"(%arg0, %arg0) {round = 0 : i1} : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
+ %11 = "tosa.arithmetic_right_shift"(%arg0, %arg0) {round = 0 : i1} : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
// CHECK: linalg.generic
// CHECK: constant 1
// CHECK: and
// CHECK: zexti
// CHECK: addi
- %11 = "tosa.arithmetic_right_shift"(%arg0, %arg0) {round = 1 : i1} : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
+ %12 = "tosa.arithmetic_right_shift"(%arg0, %arg0) {round = 1 : i1} : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
// CHECK: linalg.generic
// CHECK: cmpi
- %12 = "tosa.greater"(%0, %1) : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi1>
+ %13 = "tosa.greater"(%0, %1) : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi1>
// CHECK: linalg.generic
// CHECK: cmpi
- %13 = "tosa.greater_equal"(%0, %1) : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi1>
+ %14 = "tosa.greater_equal"(%0, %1) : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi1>
// CHECK: linalg.generic
// CHECK: select
- %14 = "tosa.select"(%12, %0, %1) : (tensor<1xi1>, tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
+ %15 = "tosa.select"(%13, %0, %1) : (tensor<1xi1>, tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
// CHECK: linalg.generic
// CHECK: cmpi
// CHECK: select
- %15 = "tosa.maximum"(%0, %1) : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
+ %16 = "tosa.maximum"(%0, %1) : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
// CHECK: linalg.generic
// CHECK: cmpi
// CHECK: select
- %16 = "tosa.minimum"(%0, %1) : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
+ %17 = "tosa.minimum"(%0, %1) : (tensor<1xi32>, tensor<1xi32>) -> tensor<1xi32>
// CHECK: linalg.generic
// CHECK: cmpi
// CHECK: select
- %17 = "tosa.clamp"(%0) {min_int = 1 : i64, max_int = 5 : i64, min_fp = 1.0 : f32, max_fp = 5.0 : f32} : (tensor<1xi32>) -> tensor<1xi32>
+ %18 = "tosa.clamp"(%0) {min_int = 1 : i64, max_int = 5 : i64, min_fp = 1.0 : f32, max_fp = 5.0 : f32} : (tensor<1xi32>) -> tensor<1xi32>
// CHECK: linalg.generic
// CHECK: cmpi
// CHECK: select
- %18 = "tosa.reluN"(%0) {max_int = 5 : i64, max_fp = 5.0 : f32} : (tensor<1xi32>) -> tensor<1xi32>
+ %19 = "tosa.reluN"(%0) {max_int = 5 : i64, max_fp = 5.0 : f32} : (tensor<1xi32>) -> tensor<1xi32>
// CHECK: linalg.generic
// CHECK: constant -32768
// CHECK: cmpi slt
// CHECK: select
// CHECK: trunci
- %19 = "tosa.cast"(%0) : (tensor<1xi32>) -> tensor<1xi16>
+ %20 = "tosa.cast"(%0) : (tensor<1xi32>) -> tensor<1xi16>
// CHECK: linalg.generic
// CHECK: yield
- %20 = "tosa.cast"(%0) : (tensor<1xi32>) -> tensor<1xi32>
+ %21 = "tosa.cast"(%0) : (tensor<1xi32>) -> tensor<1xi32>
// CHECK: linalg.generic
// CHECK: sexti
- %21 = "tosa.cast"(%0) : (tensor<1xi32>) -> tensor<1xi64>
+ %22 = "tosa.cast"(%0) : (tensor<1xi32>) -> tensor<1xi64>
// CHECK: linalg.generic
// CHECK: constant 0
// CHECK: cmpi
- %22 = "tosa.cast"(%0) : (tensor<1xi32>) -> tensor<1xi1>
+ %23 = "tosa.cast"(%0) : (tensor<1xi32>) -> tensor<1xi1>
// CHECK: linalg.generic
// CHECK: sitofp
- %23 = "tosa.cast"(%0) : (tensor<1xi32>) -> tensor<1xf32>
+ %24 = "tosa.cast"(%0) : (tensor<1xi32>) -> tensor<1xf32>
return
}