Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Switch jaxlib Python code to use the lower-level xla.ops API when bui… #2798

Merged
merged 1 commit into from
Apr 22, 2020
Merged

Switch jaxlib Python code to use the lower-level xla.ops API when bui… #2798

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 7 additions & 2 deletions jaxlib/cuda_prng.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,9 +30,14 @@

_prod = lambda xs: functools.reduce(operator.mul, xs, 1)

# TODO(phawkins): remove after we no longer need to support old jax releases.
def _unpack_builder(c):
# If `c` is a ComputationBuilder object, extracts the underlying XlaBuilder.
return getattr(c, "_builder", c)

def threefry2x32(c, keys, data):
"""ThreeFry2x32 kernel for GPU."""
c = _unpack_builder(c)
assert len(keys) == 2, keys
assert len(data) == 2, data
dims = c.GetShape(keys[0]).dimensions()
Expand All @@ -46,8 +51,8 @@ def threefry2x32(c, keys, data):
opaque = cuda_prng_kernels.cuda_threefry2x32_descriptor(_prod(dims))
layout = tuple(range(ndims - 1, -1, -1))
shape = xla_client.Shape.array_shape(dtype, dims, layout)
return c.CustomCallWithLayout(
b"cuda_threefry2x32",
return xla_client.ops.CustomCallWithLayout(
c, b"cuda_threefry2x32",
operands=(keys[0], keys[1], data[0], data[1]),
shape_with_layout=xla_client.Shape.tuple_shape([shape, shape]),
operand_shapes_with_layout=(shape,) * 4,
Expand Down
101 changes: 57 additions & 44 deletions jaxlib/cusolver.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -34,9 +34,13 @@
except ImportError:
pass


_ops = xla_client.ops
_Shape = xla_client.Shape

# TODO(phawkins): remove after we no longer need to support old jax releases.
def _unpack_builder(c):
# If `c` is a ComputationBuilder object, extracts the underlying XlaBuilder.
return getattr(c, "_builder", c)

def _real_type(dtype):
"""Returns the real equivalent of 'dtype'."""
Expand All @@ -59,6 +63,7 @@ def trsm(c, a, b, left_side=False, lower=False, trans_a=False, conj_a=False,

XLA implements unbatched triangular solve directly, so we need only implement
the batched case."""
c = _unpack_builder(c)
b_shape = c.GetShape(b)
dtype = b_shape.element_type()
dims = b_shape.dimensions()
Expand All @@ -81,8 +86,8 @@ def trsm(c, a, b, left_side=False, lower=False, trans_a=False, conj_a=False,
lwork, opaque = cublas_kernels.build_trsm_batched_descriptor(
np.dtype(dtype), batch, m, n, left_side, lower, trans_a, conj_a, diag)
layout = (num_bd, num_bd + 1) + tuple(range(num_bd - 1, -1, -1))
out = c.CustomCallWithLayout(
b"cublas_trsm_batched",
out = _ops.CustomCallWithLayout(
c, b"cublas_trsm_batched",
operands=(a, b),
shape_with_layout=_Shape.tuple_shape((
_Shape.array_shape(dtype, b_shape.dimensions(), layout),
Expand All @@ -93,11 +98,12 @@ def trsm(c, a, b, left_side=False, lower=False, trans_a=False, conj_a=False,
_Shape.array_shape(dtype, b_shape.dimensions(), layout),
),
opaque=opaque)
return c.GetTupleElement(out, 0)
return _ops.GetTupleElement(out, 0)


def potrf(c, a, lower):
"""Cholesky decomposition."""
c = _unpack_builder(c)
a_shape = c.GetShape(a)
dtype = a_shape.element_type()
dims = a_shape.dimensions()
Expand All @@ -111,8 +117,8 @@ def potrf(c, a, lower):
np.dtype(dtype), lower, batch, n)
kernel = b"cusolver_potrf"

out = c.CustomCallWithLayout(
kernel,
out = _ops.CustomCallWithLayout(
c, kernel,
operands=(a,),
shape_with_layout=_Shape.tuple_shape((
_Shape.array_shape(
Expand All @@ -126,11 +132,12 @@ def potrf(c, a, lower):
dtype, batch_dims + (n, n),
(num_bd, num_bd + 1) + tuple(range(num_bd - 1, -1, -1))),),
opaque=opaque)
return c.GetTupleElement(out, 0), c.GetTupleElement(out, 1)
return _ops.GetTupleElement(out, 0), _ops.GetTupleElement(out, 1)


def getrf(c, a):
"""LU decomposition."""
c = _unpack_builder(c)
a_shape = c.GetShape(a)
dtype = a_shape.element_type()
dims = a_shape.dimensions()
Expand All @@ -151,8 +158,8 @@ def getrf(c, a):
workspace = _Shape.array_shape(dtype, (lwork,), (0,))
kernel = b"cusolver_getrf"

out = c.CustomCallWithLayout(
kernel,
out = _ops.CustomCallWithLayout(
c, kernel,
operands=(a,),
shape_with_layout=_Shape.tuple_shape((
_Shape.array_shape(
Expand All @@ -169,11 +176,12 @@ def getrf(c, a):
dtype, batch_dims + (m, n),
(num_bd, num_bd + 1) + tuple(range(num_bd - 1, -1, -1))),),
opaque=opaque)
return (c.GetTupleElement(out, 0), c.GetTupleElement(out, 1),
c.GetTupleElement(out, 2))
return (_ops.GetTupleElement(out, 0), _ops.GetTupleElement(out, 1),
_ops.GetTupleElement(out, 2))

def geqrf(c, a):
"""QR decomposition."""
c = _unpack_builder(c)
a_shape = c.GetShape(a)
dtype = a_shape.element_type()
dims = a_shape.dimensions()
Expand All @@ -188,8 +196,8 @@ def geqrf(c, a):
workspace = _Shape.array_shape(dtype, (lwork,), (0,))
kernel = b"cusolver_geqrf"

out = c.CustomCallWithLayout(
kernel,
out = _ops.CustomCallWithLayout(
c, kernel,
operands=(a,),
shape_with_layout=_Shape.tuple_shape((
_Shape.array_shape(
Expand All @@ -206,11 +214,12 @@ def geqrf(c, a):
dtype, batch_dims + (m, n),
(num_bd, num_bd + 1) + tuple(range(num_bd - 1, -1, -1))),),
opaque=opaque)
return (c.GetTupleElement(out, 0), c.GetTupleElement(out, 1),
c.GetTupleElement(out, 2))
return (_ops.GetTupleElement(out, 0), _ops.GetTupleElement(out, 1),
_ops.GetTupleElement(out, 2))

def orgqr(c, a, tau):
"""Product of elementary Householder reflections."""
c = _unpack_builder(c)
a_shape = c.GetShape(a)
dtype = a_shape.element_type()
dims = a_shape.dimensions()
Expand All @@ -229,8 +238,8 @@ def orgqr(c, a, tau):
workspace = _Shape.array_shape(dtype, (lwork,), (0,))
kernel = b"cusolver_orgqr"

out = c.CustomCallWithLayout(
kernel,
out = _ops.CustomCallWithLayout(
c, kernel,
operands=(a, tau),
shape_with_layout=_Shape.tuple_shape((
_Shape.array_shape(
Expand All @@ -249,11 +258,12 @@ def orgqr(c, a, tau):
tuple(range(num_bd, -1, -1))),
),
opaque=opaque)
return (c.GetTupleElement(out, 0), c.GetTupleElement(out, 1))
return (_ops.GetTupleElement(out, 0), _ops.GetTupleElement(out, 1))


def syevd(c, a, lower=False):
"""Symmetric (Hermitian) eigendecomposition."""
c = _unpack_builder(c)

a_shape = c.GetShape(a)
dtype = a_shape.element_type()
Expand All @@ -276,8 +286,8 @@ def syevd(c, a, lower=False):
np.dtype(dtype), lower, batch, n)
eigvals_type = _real_type(dtype)

out = c.CustomCallWithLayout(
kernel,
out = _ops.CustomCallWithLayout(
c, kernel,
operands=(a,),
shape_with_layout=_Shape.tuple_shape((
_Shape.array_shape(dtype, dims, layout),
Expand All @@ -293,12 +303,13 @@ def syevd(c, a, lower=False):
_Shape.array_shape(dtype, dims, layout),
),
opaque=opaque)
return (c.GetTupleElement(out, 0), c.GetTupleElement(out, 1),
c.GetTupleElement(out, 2))
return (_ops.GetTupleElement(out, 0), _ops.GetTupleElement(out, 1),
_ops.GetTupleElement(out, 2))


def gesvd(c, a, full_matrices=True, compute_uv=True):
"""Singular value decomposition."""
c = _unpack_builder(c)

a_shape = c.GetShape(a)
dims = a_shape.dimensions()
Expand All @@ -316,8 +327,8 @@ def gesvd(c, a, full_matrices=True, compute_uv=True):
scalar_layout = tuple(range(num_bd - 1, -1, -1))
vector_layout = (num_bd,) + scalar_layout
matrix_layout = (num_bd, num_bd + 1) + scalar_layout
out = c.CustomCallWithLayout(
b"cusolver_gesvdj",
out = _ops.CustomCallWithLayout(
c, b"cusolver_gesvdj",
operands=(a,),
shape_with_layout=_Shape.tuple_shape((
_Shape.array_shape(dtype, batch_dims + (m, n), matrix_layout),
Expand All @@ -332,21 +343,21 @@ def gesvd(c, a, full_matrices=True, compute_uv=True):
_Shape.array_shape(dtype, batch_dims + (m, n), matrix_layout),
),
opaque=opaque)
s = c.GetTupleElement(out, 1)
u = c.GetTupleElement(out, 2)
v = c.GetTupleElement(out, 3)
info = c.GetTupleElement(out, 4)
vt = c.Transpose(v, tuple(range(num_bd)) + (num_bd + 1, num_bd))
s = _ops.GetTupleElement(out, 1)
u = _ops.GetTupleElement(out, 2)
v = _ops.GetTupleElement(out, 3)
info = _ops.GetTupleElement(out, 4)
vt = _ops.Transpose(v, tuple(range(num_bd)) + (num_bd + 1, num_bd))
if np.issubdtype(dtype, np.complexfloating):
vt = c.Conj(vt)
vt = _ops.Conj(vt)
elif m < n:
lwork, opaque = cusolver_kernels.build_gesvd_descriptor(
np.dtype(dtype), b, n, m, compute_uv, full_matrices)
scalar_layout = tuple(range(num_bd - 1, -1, -1))
vector_layout = (num_bd,) + scalar_layout
matrix_layout = (num_bd + 1, num_bd) + scalar_layout
out = c.CustomCallWithLayout(
b"cusolver_gesvd",
out = _ops.CustomCallWithLayout(
c, b"cusolver_gesvd",
operands=(a,),
shape_with_layout=_Shape.tuple_shape((
_Shape.array_shape(dtype, batch_dims + (m, n), matrix_layout),
Expand All @@ -361,19 +372,19 @@ def gesvd(c, a, full_matrices=True, compute_uv=True):
_Shape.array_shape(dtype, batch_dims + (m, n), matrix_layout),
),
opaque=opaque)
s = c.GetTupleElement(out, 1)
vt = c.GetTupleElement(out, 2)
u = c.GetTupleElement(out, 3)
info = c.GetTupleElement(out, 4)
s = _ops.GetTupleElement(out, 1)
vt = _ops.GetTupleElement(out, 2)
u = _ops.GetTupleElement(out, 3)
info = _ops.GetTupleElement(out, 4)
else:
lwork, opaque = cusolver_kernels.build_gesvd_descriptor(
np.dtype(dtype), b, m, n, compute_uv, full_matrices)

scalar_layout = tuple(range(num_bd - 1, -1, -1))
vector_layout = (num_bd,) + scalar_layout
matrix_layout = (num_bd, num_bd + 1) + scalar_layout
out = c.CustomCallWithLayout(
b"cusolver_gesvd",
out = _ops.CustomCallWithLayout(
c, b"cusolver_gesvd",
operands=(a,),
shape_with_layout=_Shape.tuple_shape((
_Shape.array_shape(dtype, batch_dims + (m, n), matrix_layout),
Expand All @@ -388,11 +399,13 @@ def gesvd(c, a, full_matrices=True, compute_uv=True):
_Shape.array_shape(dtype, batch_dims + (m, n), matrix_layout),
),
opaque=opaque)
s = c.GetTupleElement(out, 1)
u = c.GetTupleElement(out, 2)
vt = c.GetTupleElement(out, 3)
info = c.GetTupleElement(out, 4)
s = _ops.GetTupleElement(out, 1)
u = _ops.GetTupleElement(out, 2)
vt = _ops.GetTupleElement(out, 3)
info = _ops.GetTupleElement(out, 4)
if not full_matrices:
u = c.Slice(u, (0,) * len(dims), batch_dims + (m, min(m, n)))
vt = c.Slice(vt, (0,) * len(dims), batch_dims + (min(m, n), n))
u = _ops.Slice(u, (0,) * len(dims), batch_dims + (m, min(m, n)),
(1,) * len(dims))
vt = _ops.Slice(vt, (0,) * len(dims), batch_dims + (min(m, n), n),
(1,) * len(dims))
return s, u, vt, info
Loading