-
Notifications
You must be signed in to change notification settings - Fork 131
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.
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
OTFMapFusion: Minor bug fixes #1448
Conversation
Just pushed an additional fix |
if graph.out_degree(self.first_map_exit) > 1: | ||
return False | ||
if graph.in_degree(self.array) > 1: | ||
return False | ||
|
||
# Check if array is overwritten afterwards | ||
subset = next(graph.out_edges(self.first_map_exit).__iter__()).data.subset |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Would be happy to use helper functions for this. It could be less conservative.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Maybe something like is_shadowed
for an access node
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The more I think about this, the more I am convinced that the transformation should be simple and we need to write a „ShadowedAccessNode“ Pass
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Reverted the changes; We need a strategy for MapFusion, OTFMapFusion and TakletFusion. The transformations should be simple and be used with certain analysis passes
87373ae
to
ff9e2e2
Compare
This PR fixes the problem that data containers were removed from an SDFG, although they might be used in another state. Furthermore, it fixes the problem that self.second_map_entry may not point to the correct node after adding/removing nodes
This PR fixes the problem that data containers were removed from an SDFG, although they might be used in another state. Furthermore, it fixes the problem that self.second_map_entry may not point to the correct node after adding/removing nodes