And Michael Chuvelev writes:
1. In the table below the changes due to *larfp call are summarized.
I'm sorry, but your mailer mangled the table. I'm not sure which
BTW, LAWN 203 describes the changes and lists which routines are
not compatible with xLARFP:
In particular, Table 3 on page 10 points out that the -ES, -ESX,
and -EQR routines are incompatible with xLARFP, so their
auxiliary routines use xLARFG.
Table 4 on the same page lists which routines' sources were
intended to be modified. It's quite possible that I missed some,
so please send your list of unmodified routines again so I can
2. There's a lack of internal documentation (parameters description) in many
EPIR auxiliary routines (see the table below). Do you have plans to
add the description to these routines?
These all are routines for *internal* use only. The extended-
precision refinement interface is experimental and may be subject
to change, so we certainly don't want to freeze the auxiliary
Somewhere we had published a partial list of which auxiliary
routines had frozen interfaces. The rest are assumed to be
for *internal* use only and can change. It appears that list
was not propagated into newer documentation, but it was announced
with LAPACK 3.1.0. See the section on "LAPACK subroutine
interface policy" in
Yes, this needs moved into a new edition of the LUG at some
point, but many of us are busy with other writing tasks.
This e-mail and any attachments may contain confidential material for
the sole use of the intended recipient(s). Any review or distribution
by others is strictly prohibited. If you are not the intended
recipient, please contact the sender and delete all copies.
Please inform your lawyers that these trailers are going to
publicly archived mailing lists where the trailer is utterly