On Mon, 28 May 2018 23:31:57 +0200
David Adler <d.adler@xxxxxxxxx> wrote:
On 28 May 2018 19:47:46 CEST, Will Godfrey <willgodfrey@xxxxxxxxxxxxxxx>
On Mon, 28 May 2018 13:30:01 +0200 Ichthyostega <prg@xxxxxxxxxxxxxxx>...
...It seems this problem is just the tip of the iceberg and
indicates something we maybe should reconsider (after
the release of course). Can anyone recall any additional
details why these global changes in the float rounding
behaviour where introduced?
The original problem that had so worried Cal only
seemed to show up with Intel Atom N270 processors,
this is why hardly any of us could reproduce it.
It apparently resulted in a weird note pitch error.
Exactly, one of those N270s was mine. It's even still
available for tests. :-/ The notes sounded arbitrarily
off-pitch, by up to several semitones.
This email might shed light on the rationale back then.