-
Notifications
You must be signed in to change notification settings - Fork 27
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
Dimension of argument match
for matching-based scaling methods is unclear for non-square matrices
#207
Comments
Thanks @mjacobse, what a mess. It's not clear to me what the correct setting should be here. |
@nimgould any ideas on what was intended here? |
@nimgould says that the documentation is incorrect and the code is correct so it should be |
Probably copy-paste error from symmetric cases. Fixes ralna#207
Yeah that makes sense. On further inspection, the only memory issues I can find when using |
Probably copy-paste error from symmetric cases. Fixes #207
cnt counts column-wise, so should be checked up to the number of columns a%n, not the number of rows a%m. Probably copy-paste error from symmetric case. Noticed when doing tests for ralna#207 by allocating the exact sizes for each randomly generated matrix rather than the oversized maximum sizes.
cnt counts column-wise, so should be checked up to the number of columns a%n, not the number of rows a%m. Noticed when doing tests for ralna#207 by allocating the exact sizes for each randomly generated matrix rather than the oversized maximum sizes.
cnt counts column-wise, so should be checked up to the number of columns a%n, not the number of rows a%m. Noticed when doing tests for #207 by allocating the exact sizes for each randomly generated matrix rather than the oversized maximum sizes.
Documentation says
match(n)
(auction_scale_unsym, hungarian_scale_unsym), code saysmatch(m)
(auction_scale_unsym_int32, auction_scale_unsym_int64, hungarian_scale_unsym_int32, hungarian_scale_unsym_int64).The tests pass an oversized
match(maxn)
withm, n <= maxn
and the examples are using a square matrix, so they do not really provide clarification either.Indeed, neither
match(m)
normatch(n)
seem to work correctly in all cases. For both there are settings with eitherm < n
orm > n
that lead to memory access violations or uninitialized returned values.The text was updated successfully, but these errors were encountered: