Skip to content

Conversation

@therault
Copy link
Contributor

@therault therault commented Apr 4, 2025

Everything is in the title, some corner cases were not handled correctly in norms and other places when computing the numerically stable sum-square updates.

This PR fixes issue #146

Patch provided by Mathieu Faverge to import fixes done in Chameleon.

…rovided by Mathieu Faverge to import fixes done in Chameleon
@therault therault requested a review from a team as a code owner April 4, 2025 14:44
static inline void
sumsq_update( int nb, float *scale, float *sumsq, const float *value )
{
float absval = fabs(*value);

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

fabsf?

}
#endif

#endif /* _sumsq_update_h_ */

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

(Not for this PR) Why sumsq_update and sumsq_update_2?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

used in src/zlange_frb_cyclic.jdf.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

why not zsumsq_update with autoconversion

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants