cublas.py 2.39 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
# Licensed to the Apache Software Foundation (ASF) under one
# or more contributor license agreements.  See the NOTICE file
# distributed with this work for additional information
# regarding copyright ownership.  The ASF licenses this file
# to you under the Apache License, Version 2.0 (the
# "License"); you may not use this file except in compliance
# with the License.  You may obtain a copy of the License at
#
#   http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing,
# software distributed under the License is distributed on an
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND, either express or implied.  See the License for the
# specific language governing permissions and limitations
# under the License.
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48
"""External function interface to cuBLAS libraries."""
from __future__ import absolute_import as _abs

from .. import api as _api
from .. import intrin as _intrin

def matmul(lhs, rhs, transa=False, transb=False):
    """Create an extern op that compute matrix mult of A and rhs with cuBLAS

    Parameters
    ----------
    lhs : Tensor
        The left matrix operand
    rhs : Tensor
        The right matrix operand
    transa : bool
        Whether transpose lhs
    transb : bool
        Whether transpose rhs

    Returns
    -------
    C : Tensor
        The result tensor.
    """
    n = lhs.shape[1] if transa else lhs.shape[0]
    m = rhs.shape[0] if transb else rhs.shape[1]
    return _api.extern(
        (n, m), [lhs, rhs],
        lambda ins, outs: _intrin.call_packed(
            "tvm.contrib.cublas.matmul",
            ins[0], ins[1], outs[0], transa, transb), name="C")
49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76

def batch_matmul(lhs, rhs, transa=False, transb=False):
    """Create an extern op that compute batch matrix mult of A and rhs with cuBLAS

    Parameters
    ----------
    lhs : Tensor
        The left matrix operand
    rhs : Tensor
        The right matrix operand
    transa : bool
        Whether transpose lhs
    transb : bool
        Whether transpose rhs

    Returns
    -------
    C : Tensor
        The result tensor.
    """
    b = lhs.shape[0]
    n = lhs.shape[2] if transa else lhs.shape[1]
    m = rhs.shape[1] if transb else rhs.shape[2]
    return _api.extern(
        (b, n, m), [lhs, rhs],
        lambda ins, outs: _intrin.call_packed(
            "tvm.contrib.cublas.batch_matmul",
            ins[0], ins[1], outs[0], transa, transb), name="C")