Fix integer overflow in TensorProductMatrix #5911
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
/run-tests |
kronbichler
reviewed
Feb 17, 2018
| } | ||
| return n_rows_max; | ||
| }; | ||
| const unsigned int nm_flat_size_max = n_rows_max() * n_rows_max() * macro_size; |
There was a problem hiding this comment.
I would not make this a lambda (without argument) since you just count the components. You can simply do this inline for the variable n_rows_max.
| std::array<std::size_t,dim> n_rows; | ||
| for (unsigned int d = 0; d < dim; ++d) | ||
| n_rows[d] = mass_matrix[d].n_rows(); | ||
| n_rows_max = *(std::max_element (n_rows.cbegin(),n_rows.cend())); |
There was a problem hiding this comment.
This looks too complicated. Why not simply do the computation as
std::size_t n_rows_max = 0;
for (unsigned int d=0; d<dim; ++d)
n_rows_max = std::max(n_rows_max, mass_matrix[d].n_rows();
There was a problem hiding this comment.
Yes, thanks! In addition, I replaced the reserve with resize and deleted redundant resizes inside the loop, since vectorized_transpose_and_store has the control over the actual data size.
jwitte08
force-pushed
the
int_overflow_tpmatrix
branch
2 times, most recently
from
a4a8fc3
to
c1da6eb
Compare
jwitte08
force-pushed
the
int_overflow_tpmatrix
branch
from
c1da6eb
to
c42a128
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I have encountered an integer overflow for a 3D use case with compile-time constant size=144. The guilty part causing the overflow is
fixed_int_power, where the guilty part is actually me!What we really want to do is to determine the maximum row-size of all directional mass matrices and use this to reserve memory for our (intermediate) flat data arrays. This avoids a re-allocation within the loop at line 758. Mistakenly, I took the
dimth power of the row-sizes.In my opinion, it makes sense to avoid a re-allocation, but maybe it is just overly complicated and one should just do the
resizes in the loop. Note that a re-allocation would only appear in the case where we have a non-isotropic Kronecker decomposition.I am open to any suggestion!