Skip to content

Test async changes #1298

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
merged 2 commits into from
Jul 21, 2023
Merged

Test async changes #1298

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
de1b1a4
test_async_submit to use input array with size a multiple of max_work…
oleksandr-pavlyk Jul 19, 2023
22694b5
Changes to test_async_submit
oleksandr-pavlyk Jul 21, 2023
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
38 changes: 23 additions & 15 deletions dpctl/tests/test_sycl_kernel_submit.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -167,16 +167,21 @@ def test_async_submit():
assert isinstance(kern2Kernel, dpctl_prog.SyclKernel)

status_complete = dpctl.event_status_type.complete
n = 256 * 1024
X = dpt.empty((3, n), dtype="u4", usm_type="device", sycl_queue=q)

# choose input size based on capability of the device
f = q.sycl_device.max_work_group_size
n = f * 1024
n_alloc = 4 * n

X = dpt.empty((3, n_alloc), dtype="u4", usm_type="device", sycl_queue=q)
first_row = dpctl_mem.as_usm_memory(X[0])
second_row = dpctl_mem.as_usm_memory(X[1])
third_row = dpctl_mem.as_usm_memory(X[2])

p1, p2 = 17, 27

async_detected = False
for _ in range(5):
for attempt in range(5):
e1 = q.submit(
kern1Kernel,
[
Expand Down Expand Up @@ -209,19 +214,22 @@ def test_async_submit():
e3_st = e3.execution_status
e2_st = e2.execution_status
e1_st = e1.execution_status
if not all(
[
e == status_complete
for e in (
e1_st,
e2_st,
e3_st,
)
]
):
are_complete = [
e == status_complete
for e in (
e1_st,
e2_st,
e3_st,
)
]
e3.wait()
if not all(are_complete):
async_detected = True
e3.wait()
break
else:
n = n * (1 if attempt % 2 == 0 else 2)
if n > n_alloc:
break

assert async_detected, "No evidence of async submission detected, unlucky?"
Xnp = dpt.asnumpy(X)
Expand All @@ -231,4 +239,4 @@ def test_async_submit():
Xref[1, i] = (i * i * i) % p2
Xref[2, i] = min(Xref[0, i], Xref[1, i])

assert np.array_equal(Xnp, Xref)
assert np.array_equal(Xnp[:, :n], Xref[:, :n])