Skip to content

compiler: Enable blocking-before-CIRE#2948

Open
FabioLuporini wants to merge 10 commits into
mainfrom
smarter-tuner-3
Open

compiler: Enable blocking-before-CIRE#2948
FabioLuporini wants to merge 10 commits into
mainfrom
smarter-tuner-3

Conversation

@FabioLuporini

Copy link
Copy Markdown
Contributor

This is key for features in PRO

In doing so, several utility functions are added, which might come in handy in the future as well

Some clean up also, which is never a bad thing

else:
self.umt_reduce = UnboundTuple(*par_tile.default, 1)

def next(self, prefix, d, clusters):

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

significant reduction in complexity

Comment thread tests/test_dle.py
assert iters[0].step == par_tile[1]
assert iters[1].step == par_tile[0]

def test_custom_rule0(self):

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

unused feature, dropped; so dropping the tests too

@codecov

codecov Bot commented Jun 16, 2026

Copy link
Copy Markdown

Codecov Report

❌ Patch coverage is 88.20225% with 42 lines in your changes missing coverage. Please review.
✅ Project coverage is 79.40%. Comparing base (58df797) to head (b340037).
⚠️ Report is 2 commits behind head on main.

Files with missing lines Patch % Lines
devito/passes/clusters/aliases.py 83.59% 20 Missing and 1 partial ⚠️
devito/passes/clusters/blocking.py 90.19% 3 Missing and 2 partials ⚠️
devito/ir/support/guards.py 73.33% 3 Missing and 1 partial ⚠️
devito/types/parallel.py 66.66% 4 Missing ⚠️
devito/ir/clusters/cluster.py 86.36% 2 Missing and 1 partial ⚠️
devito/passes/iet/engine.py 0.00% 3 Missing ⚠️
devito/core/operator.py 50.00% 1 Missing and 1 partial ⚠️
Additional details and impacted files
@@            Coverage Diff             @@
##             main    #2948      +/-   ##
==========================================
- Coverage   83.47%   79.40%   -4.08%     
==========================================
  Files         249      249              
  Lines       52276    52436     +160     
  Branches     4503     4512       +9     
==========================================
- Hits        43638    41635    -2003     
- Misses       7880     9988    +2108     
- Partials      758      813      +55     
Flag Coverage Δ
pytest-gpu-aomp-amdgpuX ?
pytest-gpu-gcc- 78.25% <88.20%> (+0.03%) ⬆️
pytest-gpu-icx- 78.17% <88.20%> (-0.01%) ⬇️
pytest-gpu-nvc-nvidiaX ?

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

"""Return the DataSpace of this ClusterGroup."""
return DataSpace.union(*[i.dspace.reset() for i in self])

@property

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

cached?

Comment thread devito/core/gpu.py

# Overrides the default values in the main Operator class
BLOCK_LEVELS = 0
CIRE_BLOCK_TEMPS = False

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

Is it intended to be a default or enforced?

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

enforced or things would break

if not mapper:
return self

if self.halo_scheme:

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

Can probably be relaxed to self.halo_scheme.distributed_aindices & mapper.keys()

# Handle ComponentAccesses
subs = {i: Symbol(f'dummy{n}') for n, i in enumerate(compaccs)}
exprs1 = uxreplace(exprs, subs)
terms1 = retrieve_terminals(exprs1, **kwargs)

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

would likely be "cheaper" to have an ignore=compaccs in retrieve_terminals

@review-notebook-app

Copy link
Copy Markdown

Check out this pull request on  ReviewNB

See visual diffs & provide feedback on Jupyter Notebooks.


Powered by ReviewNB

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants