RuntimeError: BinaryOpType cannot be mapped to BcastOpMath

[meenakshiramanathan1]

Description

• Floor op faces this runtime error

Reproduce

git checkout  mramanathan/grid_sample_floor_issue
pytest forge/test/mlir/test_ops.py::test_floor

Observed Behaviour

RuntimeError: TT_THROW @ /proj_sw/user_dev/mramanathan/bgdlab18_dec24_forge/tt-forge-fe/third_party/tt-mlir/third_party/tt-metal/src/tt-metal/ttnn/cpp/ttnn/operations/eltwise/binary/device/broadcast_height_and_width_multi_core_program_factory.cpp:27: tt::exception
E       info:
E       BinaryOpType cannot be mapped to BcastOpMath

Logs

• floor.log

TTIR graph

module @Floor attributes {tt.system_desc = #tt.system_desc<[{role = host, target_triple = "x86_64-pc-linux-gnu"}], [{arch = <wormhole_b0>, grid = 8x8, l1_size = 1499136, num_dram_channels = 12, dram_channel_size = 1073741824, noc_l1_address_align_bytes = 16, pcie_address_align_bytes = 32, noc_dram_address_align_bytes = 32, l1_unreserved_base = 1024, erisc_l1_unreserved_base = 1024, dram_unreserved_base = 1024, dram_unreserved_end = 1073741824, physical_cores = {worker = [ 0x0,  0x1,  0x2,  0x3,  0x4,  0x5,  0x6,  0x7,  1x0,  1x1,  1x2,  1x3,  1x4,  1x5,  1x6,  1x7,  2x0,  2x1,  2x2,  2x3,  2x4,  2x5,  2x6,  2x7,  3x0,  3x1,  3x2,  3x3,  3x4,  3x5,  3x6,  3x7,  4x0,  4x1,  4x2,  4x3,  4x4,  4x5,  4x6,  4x7,  5x0,  5x1,  5x2,  5x3,  5x4,  5x5,  5x6,  5x7,  6x0,  6x1,  6x2,  6x3,  6x4,  6x5,  6x6,  6x7,  7x0,  7x1,  7x2,  7x3,  7x4,  7x5,  7x6,  7x7] dram = [ 8x0,  9x0,  10x0,  8x1,  9x1,  10x1,  8x2,  9x2,  10x2,  8x3,  9x3,  10x3]}, supported_data_types = [<f32>, <f16>, <bf16>, <bfp_f8>, <bfp_bf8>, <bfp_f4>, <bfp_bf4>, <bfp_f2>, <bfp_bf2>, <u32>, <u16>, <u8>], supported_tile_sizes = [ 4x16,  16x16,  32x16,  4x32,  16x32,  32x32], num_cbs = 32}], [0], [3 : i32], [ 0x0x0x0]>} {
  func.func @forward(%arg0: tensor<6xf32> {ttir.name = "a"}, %arg1: tensor<1xf32> {ttir.name = "input_1_less_26"}, %arg2: tensor<1xf32> {ttir.name = "dc.input_tensor.where_30.0"}, %arg3: tensor<1xf32> {ttir.name = "input_2_where_30"}) -> (tensor<6xf32> {ttir.name = "Floor_3.output_subtract_31"}) {
    %0 = tensor.empty() : tensor<6xi32>
    %1 = "ttir.typecast"(%arg0, %0) <{operandSegmentSizes = array<i32: 1, 1>}> {dtype = "Int32"} : (tensor<6xf32>, tensor<6xi32>) -> tensor<6xi32>
    %2 = tensor.empty() : tensor<6xf32>
    %3 = "ttir.typecast"(%1, %2) <{operandSegmentSizes = array<i32: 1, 1>}> {dtype = "Float32"} : (tensor<6xi32>, tensor<6xf32>) -> tensor<6xf32>
    %4 = tensor.empty() : tensor<6xf32>
    %5 = "ttir.le"(%arg0, %arg1, %4) <{operandSegmentSizes = array<i32: 2, 1>}> : (tensor<6xf32>, tensor<1xf32>, tensor<6xf32>) -> tensor<6xf32>
    %6 = tensor.empty() : tensor<6xi8>
    %7 = "ttir.typecast"(%5, %6) <{operandSegmentSizes = array<i32: 1, 1>}> {dtype = "Int8"} : (tensor<6xf32>, tensor<6xi8>) -> tensor<6xi8>
    %8 = tensor.empty() : tensor<6xf32>
    %9 = "ttir.ne"(%arg0, %3, %8) <{operandSegmentSizes = array<i32: 2, 1>}> : (tensor<6xf32>, tensor<6xf32>, tensor<6xf32>) -> tensor<6xf32>
    %10 = tensor.empty() : tensor<6xf32>
    %11 = "ttir.typecast"(%9, %10) <{operandSegmentSizes = array<i32: 1, 1>}> {dtype = "Float32"} : (tensor<6xf32>, tensor<6xf32>) -> tensor<6xf32>
    %12 = tensor.empty() : tensor<6xf32>
    %13 = "ttir.multiply"(%7, %11, %12) <{operandSegmentSizes = array<i32: 2, 1>}> : (tensor<6xi8>, tensor<6xf32>, tensor<6xf32>) -> tensor<6xf32>
    %14 = tensor.empty() : tensor<6xf32>
    %15 = "ttir.subtract"(%arg2, %7, %14) <{operandSegmentSizes = array<i32: 2, 1>}> : (tensor<1xf32>, tensor<6xi8>, tensor<6xf32>) -> tensor<6xf32>
    %16 = tensor.empty() : tensor<6xf32>
    %17 = "ttir.multiply"(%15, %arg3, %16) <{operandSegmentSizes = array<i32: 2, 1>}> : (tensor<6xf32>, tensor<1xf32>, tensor<6xf32>) -> tensor<6xf32>
    %18 = tensor.empty() : tensor<6xf32>
    %19 = "ttir.add"(%13, %17, %18) <{operandSegmentSizes = array<i32: 2, 1>}> : (tensor<6xf32>, tensor<6xf32>, tensor<6xf32>) -> tensor<6xf32>
    %20 = tensor.empty() : tensor<6xf32>
    %21 = "ttir.subtract"(%3, %19, %20) <{operandSegmentSizes = array<i32: 2, 1>}> : (tensor<6xf32>, tensor<6xf32>, tensor<6xf32>) -> tensor<6xf32>
    return %21 : tensor<6xf32>
  }
}