Skip to content

normalize.py: Use count_nonzero in place of the reduction kernels #265

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

Open
mfep wants to merge 1 commit into
base: main
Choose a base branch
from
Open

normalize.py: Use count_nonzero in place of the reduction kernels #265

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
24 changes: 2 additions & 22 deletions httomolibgpu/prep/normalize.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -152,26 +152,6 @@ def dark_flat_field_correction(
no_return=True,
)

count_greater_kernel = cp.ReductionKernel(
in_params="T data, raw float32 upper_bound",
out_params="int32 out",
map_expr="data >= upper_bound ? 1 : 0", # map each element → 1 or 0
reduce_expr="a + b", # sum them
post_map_expr="out = a", # final result
identity="0",
name="count_greater",
)

count_smaller_kernel = cp.ReductionKernel(
in_params="T data, raw float32 lower_bound",
out_params="int32 out",
map_expr="data <= lower_bound ? 1 : 0", # map each element → 1 or 0
reduce_expr="a + b", # sum them
post_map_expr="out = a", # final result
identity="0",
name="count_smaller",
)

normalisation_kernel(data, flat0, dark0, upper_bound, lower_bound, out)

if clipping_warning:
Expand All @@ -180,9 +160,9 @@ def dark_flat_field_correction(
50.0 # warning if more clipped values than given percentage
)

clipped_total_up = int(count_greater_kernel(out, float32(upper_bound)))
clipped_total_up = cp.count_nonzero(out >= upper_bound)

@dkazanc dkazanc Sep 7, 2026

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I remember with CuPy 12.3 I started with this in-built function but it was performing poorly. I guess it has been improved. Will test.

clipped_up_percent = clipped_total_up / data_elements_num * 100
clipped_total_lower = int(count_smaller_kernel(out, float32(lower_bound)))
clipped_total_lower = cp.count_nonzero(out <= lower_bound)
clipped_down_percent = clipped_total_lower / data_elements_num * 100

if clipped_up_percent >= clipped_percentage_warning:
Expand Down
Loading