You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In 32-bit OS with sizeof(int) == sizeof(long) == 4, binop_with_scalar fails with the following message: binop_with_scalar.log.
Other 31 tests in accumulators passed successfully.
The text was updated successfully, but these errors were encountered:
The test is relatively new (for issue #79). The error is only with float accumulators: discrepancy between vector and scalar float accumulators' behavior.
We saw something similar before: #181 (comment) .
Should we simply increase the tolerances to 1E-3, although this looks too much, or shall we dig further?
...and added a test.
The handling of the situation is still not ideal: if declaring a named
accumulator holding `vector<float>` is attempted, compilation fails
with `static_assert` issuing a relevant message, but other compilation
errors are triggered that add clutter.
However, this should close#587.
Cross-ref: #181 (comment)
In 32-bit OS with sizeof(int) == sizeof(long) == 4, binop_with_scalar fails with the following message: binop_with_scalar.log.
Other 31 tests in accumulators passed successfully.
The text was updated successfully, but these errors were encountered: