def default_partition_fn(keys, shard_num):
"""The default partition function.
partition keys by "mod" strategy.
keys: a tensor presents the keys to be partitioned.
shard_num: the num of partitions
Returns:
a tensor with same shape as keys with type of `tf.int32`,
represents the corresponding partition-ids of keys.
"""
keys_op = ops.convert_to_tensor(keys, name="keys")
gpu_mode = _pywrap_util_port.IsGoogleCudaEnabled()
with ops.colocate_with(keys_op):
if keys_op.dtype == dtypes.int64 and gpu_mode:
# This branch has low performance on some multi-CPU scenario,
# so we try to use default branch when GPUs are not available.
mask = constant_op.constant(0x7fffffff, dtypes.int64)
keys_int32 = math_ops.cast(bitwise_ops.bitwise_and(keys_op, mask),
dtypes.int32)
mod = math_ops.mod(keys_int32,
constant_op.constant(shard_num, dtypes.int32))
ids = math_ops.cast(mod, dtype=dtypes.int32)
elif keys_op.dtype == dtypes.string:
ids = string_ops.string_to_hash_bucket_fast(keys_op, shard_num)
mask = constant_op.constant(0x7fffffff, dtypes.int64)
ids = math_ops.cast(bitwise_ops.bitwise_and(ids, mask),
dtypes.int32)
else:
ids = math_ops.cast(math_ops.mod(keys_op, shard_num),
dtype=dtypes.int32)
return ids
def testInvertOp(self):
dtype_list = [
dtypes.int8, dtypes.int16, dtypes.int32, dtypes.int64,
dtypes.uint8, dtypes.uint16, dtypes.uint32, dtypes.uint64
]
inputs = [0, 5, 3, 14]
with self.test_session(use_gpu=True) as sess:
for dtype in dtype_list:
# Because of issues with negative numbers, let's test this indirectly.
# 1. invert(a) and a = 0
# 2. invert(a) or a = invert(0)
input_tensor = constant_op.constant(inputs, dtype=dtype)
not_a_and_a, not_a_or_a, not_0 = sess.run([
bitwise_ops.bitwise_and(input_tensor,
bitwise_ops.invert(input_tensor)),
bitwise_ops.bitwise_or(input_tensor,
bitwise_ops.invert(input_tensor)),
bitwise_ops.invert(constant_op.constant(0, dtype=dtype))
])
self.assertAllEqual(not_a_and_a, [0, 0, 0, 0])
self.assertAllEqual(not_a_or_a, [not_0] * 4)
# For unsigned dtypes let's also check the result directly.
if dtype.is_unsigned:
inverted = sess.run(bitwise_ops.invert(input_tensor))
expected = [dtype.max - x for x in inputs]
self.assertAllEqual(inverted, expected)
def testShapeInference(self):
dtype_list = [dtypes.int8, dtypes.int16, dtypes.int32, dtypes.int64,
dtypes.uint8, dtypes.uint16]
with self.test_session(use_gpu=True) as sess:
for dtype in dtype_list:
lhs = constant_op.constant([[0], [3], [5]], dtype=dtype)
rhs = constant_op.constant([[1, 2, 4]], dtype=dtype)
and_tensor = bitwise_ops.bitwise_and(lhs, rhs)
or_tensor = bitwise_ops.bitwise_or(lhs, rhs)
xor_tensor = bitwise_ops.bitwise_xor(lhs, rhs)
ls_tensor = bitwise_ops.left_shift(lhs, rhs)
rs_tensor = bitwise_ops.right_shift(lhs, rhs)
and_result, or_result, xor_result, ls_result, rs_result = sess.run(
[and_tensor, or_tensor, xor_tensor, ls_tensor, rs_tensor])
# Compare shape inference with result
self.assertAllEqual(and_tensor.get_shape().as_list(), and_result.shape)
self.assertAllEqual(and_tensor.get_shape().as_list(), [3, 3])
self.assertAllEqual(or_tensor.get_shape().as_list(), or_result.shape)
self.assertAllEqual(or_tensor.get_shape().as_list(), [3, 3])
self.assertAllEqual(xor_tensor.get_shape().as_list(), xor_result.shape)
self.assertAllEqual(xor_tensor.get_shape().as_list(), [3, 3])
self.assertAllEqual(ls_tensor.get_shape().as_list(), ls_result.shape)
self.assertAllEqual(ls_tensor.get_shape().as_list(), [3, 3])
self.assertAllEqual(rs_tensor.get_shape().as_list(), rs_result.shape)
self.assertAllEqual(rs_tensor.get_shape().as_list(), [3, 3])
def testShapeInference(self):
dtype_list = [dtypes.int8, dtypes.int16, dtypes.int32, dtypes.int64,
dtypes.uint8, dtypes.uint16]
with self.session(use_gpu=True) as sess:
for dtype in dtype_list:
lhs = constant_op.constant([[0], [3], [5]], dtype=dtype)
rhs = constant_op.constant([[1, 2, 4]], dtype=dtype)
and_tensor = bitwise_ops.bitwise_and(lhs, rhs)
or_tensor = bitwise_ops.bitwise_or(lhs, rhs)
xor_tensor = bitwise_ops.bitwise_xor(lhs, rhs)
ls_tensor = bitwise_ops.left_shift(lhs, rhs)
rs_tensor = bitwise_ops.right_shift(lhs, rhs)
and_result, or_result, xor_result, ls_result, rs_result = sess.run(
[and_tensor, or_tensor, xor_tensor, ls_tensor, rs_tensor])
# Compare shape inference with result
self.assertAllEqual(and_tensor.get_shape().as_list(), and_result.shape)
self.assertAllEqual(and_tensor.get_shape().as_list(), [3, 3])
self.assertAllEqual(or_tensor.get_shape().as_list(), or_result.shape)
self.assertAllEqual(or_tensor.get_shape().as_list(), [3, 3])
self.assertAllEqual(xor_tensor.get_shape().as_list(), xor_result.shape)
self.assertAllEqual(xor_tensor.get_shape().as_list(), [3, 3])
self.assertAllEqual(ls_tensor.get_shape().as_list(), ls_result.shape)
self.assertAllEqual(ls_tensor.get_shape().as_list(), [3, 3])
self.assertAllEqual(rs_tensor.get_shape().as_list(), rs_result.shape)
self.assertAllEqual(rs_tensor.get_shape().as_list(), [3, 3])
def _benchmark_bitwise_and(self, mat, device):
if device == GPU and not context.num_gpus():
return
with context.device(device):
if device == GPU:
mat = mat.gpu()
func = lambda: bitwise_ops.bitwise_and(mat, mat)
self._run(func, num_iters=5000)
def testBinaryOps(self):
dtype_list = [dtypes.int8, dtypes.int16, dtypes.int32, dtypes.int64,
dtypes.uint8, dtypes.uint16]
with self.test_session(use_gpu=True) as sess:
for dtype in dtype_list:
lhs = constant_op.constant([0, 5, 3, 14], dtype=dtype)
rhs = constant_op.constant([5, 0, 7, 11], dtype=dtype)
and_result, or_result, xor_result = sess.run(
[bitwise_ops.bitwise_and(lhs, rhs),
bitwise_ops.bitwise_or(lhs, rhs),
bitwise_ops.bitwise_xor(lhs, rhs)])
self.assertAllEqual(and_result, [0, 0, 3, 10])
self.assertAllEqual(or_result, [5, 5, 7, 15])
self.assertAllEqual(xor_result, [5, 5, 4, 5])
def testBinaryOps(self):
dtype_list = [dtypes.int8, dtypes.int16, dtypes.int32, dtypes.int64,
dtypes.uint8, dtypes.uint16, dtypes.uint32, dtypes.uint64]
with self.session(use_gpu=True) as sess:
for dtype in dtype_list:
lhs = constant_op.constant([0, 5, 3, 14], dtype=dtype)
rhs = constant_op.constant([5, 0, 7, 11], dtype=dtype)
and_result, or_result, xor_result = sess.run(
[bitwise_ops.bitwise_and(lhs, rhs),
bitwise_ops.bitwise_or(lhs, rhs),
bitwise_ops.bitwise_xor(lhs, rhs)])
self.assertAllEqual(and_result, [0, 0, 3, 10])
self.assertAllEqual(or_result, [5, 5, 7, 15])
self.assertAllEqual(xor_result, [5, 5, 4, 5])
def default_partition_fn(keys, shard_num):
"""The default partition function.
partition keys by "mod" strategy.
keys: a tensor presents the keys to be partitioned.
shard_num: the num of partitions
Returns:
a tensor with same shape as keys with type of `tf.int32`,
represents the corresponding partition-ids of keys.
"""
keys_op = ops.convert_to_tensor(keys, name="keys")
with ops.colocate_with(keys_op):
if keys_op.dtype == dtypes.int64:
mask = constant_op.constant(0x7FFFFFFF, dtypes.int64)
keys_int32 = math_ops.cast(bitwise_ops.bitwise_and(keys_op, mask),
dtypes.int32)
mod = math_ops.mod(keys_int32,
constant_op.constant(shard_num, dtypes.int32))
ids = math_ops.cast(mod, dtype=dtypes.int32)
else:
ids = math_ops.cast(math_ops.mod(keys_op, shard_num), dtype=dtypes.int32)
return ids
def testInvertOp(self):
dtype_list = [dtypes.int8, dtypes.int16, dtypes.int32, dtypes.int64,
dtypes.uint8, dtypes.uint16]
inputs = [0, 5, 3, 14]
with self.test_session(use_gpu=True) as sess:
for dtype in dtype_list:
# Because of issues with negative numbers, let's test this indirectly.
# 1. invert(a) and a = 0
# 2. invert(a) or a = invert(0)
input_tensor = constant_op.constant(inputs, dtype=dtype)
not_a_and_a, not_a_or_a, not_0 = sess.run(
[bitwise_ops.bitwise_and(
input_tensor, bitwise_ops.invert(input_tensor)),
bitwise_ops.bitwise_or(
input_tensor, bitwise_ops.invert(input_tensor)),
bitwise_ops.invert(constant_op.constant(0, dtype=dtype))])
self.assertAllEqual(not_a_and_a, [0, 0, 0, 0])
self.assertAllEqual(not_a_or_a, [not_0] * 4)
# For unsigned dtypes let's also check the result directly.
if dtype.is_unsigned:
inverted = sess.run(bitwise_ops.invert(input_tensor))
expected = [dtype.max - x for x in inputs]
self.assertAllEqual(inverted, expected)