Skip to content

Adopt linalg functions from LAPACK extension to asynchronous dpctl execution #1922

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

Merged
antonwolfy merged 6 commits into from
Jul 15, 2024
Merged

Adopt linalg functions from LAPACK extension to asynchronous dpctl execution #1922

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
4261cf2
Update LinAlg functions from BLAS routine
antonwolfy Jul 12, 2024
06911f6
Decouple einsum utils function to separate file
antonwolfy Jul 12, 2024
48d3ecf
Update LinAlg functions from LAPACK routine
antonwolfy Jul 12, 2024
90998a0
Merge branch 'master' into adopt-lapack-to-async-dpctl-execution
antonwolfy Jul 13, 2024
dd58ae2
Removed batch_call from dpnp_svd()
antonwolfy Jul 15, 2024
f38dabd
Merge branch 'master' into adopt-lapack-to-async-dpctl-execution
antonwolfy Jul 15, 2024
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
25 changes: 13 additions & 12 deletions dpnp/dpnp_utils/dpnp_utils_linearalgebra.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -116,15 +116,14 @@ def _copy_array(x, copy_flag=False, dtype=None, order="C"):

exec_q = x_copy.sycl_queue
_manager = dpu.SequentialOrderManager[exec_q]
dep_evs = _manager.submitted_events

ht_copy_ev, copy_ev = ti._copy_usm_ndarray_into_usm_ndarray(
ht_ev, copy_ev = ti._copy_usm_ndarray_into_usm_ndarray(
src=dpnp.get_usm_ndarray(x),
dst=x_copy.get_array(),
sycl_queue=exec_q,
depends=dep_evs,
depends=_manager.submitted_events,
)
_manager.add_event_pair(ht_copy_ev, copy_ev)
_manager.add_event_pair(ht_ev, copy_ev)
return x_copy
return x

Expand Down Expand Up @@ -356,14 +355,14 @@ def _gemm_batch_matmul(exec_q, x1, x2, res):
x2_usm = dpnp.get_usm_ndarray(x2[i : i + chunk, ...])
res_usm = dpnp.get_usm_ndarray(res[i : i + chunk, ...])

ht_blas_ev, blas_ev, row_major = bi._gemm_batch(
ht_ev, blas_ev, row_major = bi._gemm_batch(
exec_q,
x1_usm,
x2_usm,
res_usm,
depends=_manager.submitted_events,
)
_manager.add_event_pair(ht_blas_ev, blas_ev)
_manager.add_event_pair(ht_ev, blas_ev)

res_shape = res.shape
_, res_is_c_contig, res_is_f_contig = _define_contig_flag(res)
Expand All @@ -388,14 +387,15 @@ def _gemm_batch_matmul(exec_q, x1, x2, res):

def _gemm_matmul(exec_q, x1, x2, res):
_manager = dpu.SequentialOrderManager[exec_q]
ht_gemm_ev, gemm_ev, row_major = bi._gemm(

ht_ev, gemm_ev, row_major = bi._gemm(
exec_q,
dpnp.get_usm_ndarray(x1),
dpnp.get_usm_ndarray(x2),
dpnp.get_usm_ndarray(res),
depends=_manager.submitted_events,
)
_manager.add_event_pair(ht_gemm_ev, gemm_ev)
_manager.add_event_pair(ht_ev, gemm_ev)

if row_major:
if res.flags.f_contiguous is True:
Expand Down Expand Up @@ -635,14 +635,14 @@ def dpnp_dot(a, b, /, out=None, *, conjugate=False):
else:
dot_func = "_dot"

ht_dot_ev, dot_ev = getattr(bi, dot_func)(
ht_ev, dot_ev = getattr(bi, dot_func)(
exec_q,
dpnp.get_usm_ndarray(a),
dpnp.get_usm_ndarray(b),
dpnp.get_usm_ndarray(result),
depends=_manager.submitted_events,
)
_manager.add_event_pair(ht_dot_ev, dot_ev)
_manager.add_event_pair(ht_ev, dot_ev)
else:
# oneapi::mkl::blas::dot is slow for integer data type,
# so using dpctl.tensor.vecdot instead
Expand Down Expand Up @@ -866,15 +866,16 @@ def dpnp_matmul(
x_usm = dpnp.get_usm_ndarray(x2)

_manager = dpu.SequentialOrderManager[exec_q]
ht_gemv_ev, gemv_ev = bi._gemv(

ht_ev, gemv_ev = bi._gemv(
exec_q,
a_usm,
x_usm,
dpnp.get_usm_ndarray(result),
transpose,
depends=_manager.submitted_events,
)
_manager.add_event_pair(ht_gemv_ev, gemv_ev)
_manager.add_event_pair(ht_ev, gemv_ev)
elif call_flag == "gemm":
result = _gemm_matmul(
exec_q,
Expand Down
Loading