Open seanm opened 4 weeks ago
@blowekamp maybe this is similar to https://github.com/InsightSoftwareConsortium/ITK/issues/3031 ?
The other was a logic bug in the algorithm, which required a good amount of effort in figuring out details of the method. This is a different algorithm.
I could look into the SLIC TSan issues, as I contributed that class.
I could look into the SLIC TSan issues, as I contributed that class.
I triaged them last night, and will write up a ticket for them shortly.
Running the itkParallelSparseFieldLevelSetImageFilterTest test with TSan:
We see from the above that one thread is writing to 0x00010b524516 and another is simultaneously reading from the same address.
Here is the relevant code, NB the fire emojis:
To me, this indeed looks like a buggy race. But I don't know this code at all.