<div dir="ltr">Committed with additional tests and comments as requested in r<span style="font-size:13px">225123.</span><div><span style="font-size:13px"><br></span></div><div><span style="font-size:13px">Thanks for the review Mehdi.</span></div><div><span style="font-size:13px"><br></span></div><div><span style="font-size:13px">- Lang.</span></div></div><div class="gmail_extra"><br><div class="gmail_quote">On Fri, Jan 2, 2015 at 7:26 PM, Mehdi Amini <span dir="ltr"><<a href="mailto:mehdi.amini@apple.com" target="_blank">mehdi.amini@apple.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div style="word-wrap:break-word">OK I see now, your change enforce the “(exactly)” in the comment.<div><br></div><div>The commit LGTM.</div><div><br></div><div>Some comments:</div><div><br></div><div>I would still love to see a comment in the unit test file (as it is done earlier in the file for the other tests).</div><div><br></div><div>Also the unit test does not test the <span style="font-family:monospace,monospace">rmTowardNegative</span> case. It seems to be never tested by the validation. I changed</div><span class=""><div><span style="font-family:monospace,monospace"><br></span></div><div><span style="font-family:monospace,monospace">sign = (rounding_mode == rmTowardNegative);</span></div><div><span style="font-family:monospace,monospace"><br></span></div></span><div>for </div><div><br></div><div><span style="font-family:monospace,monospace">sign = 0; </span></div><div><span style="font-family:monospace,monospace"><br></span></div><div>and ninja check-all is still passing.</div><span class="HOEnZb"><font color="#888888"><div><br></div><div>Mehdi</div></font></span><div><div class="h5"><div><br></div><div><br></div><div><div><blockquote type="cite"><div>On Jan 2, 2015, at 4:59 PM, Lang Hames <<a href="mailto:lhames@gmail.com" target="_blank">lhames@gmail.com</a>> wrote:</div><br><div><div dir="ltr">Hi Mehdi,<div><br></div><div>Thanks for the feedback. I think the comment is fine - the problem is just that the original conditional didn't exactly match it. With the addition of the underflow check I believe the condition now matches the comment: The sign is only tweaked if the result of the addition is exactly zero, rather than a small result that truncates to zero.</div><div><br></div><div>CC'ing llvm-commits, where this email should have gone in the first place. :)<div><br></div></div><div>Cheers,</div><div>Lang.</div></div><div class="gmail_extra"><br><div class="gmail_quote">On Thu, Jan 1, 2015 at 10:56 PM, Mehdi Amini <span dir="ltr"><<a href="mailto:mehdi.amini@apple.com" target="_blank">mehdi.amini@apple.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div style="word-wrap:break-word">Hi Lang,<div><br><div><div><div><blockquote type="cite"><div>On Jan 1, 2015, at 5:57 PM, Lang Hames <<a href="mailto:lhames@gmail.com" target="_blank">lhames@gmail.com</a>> wrote:</div><br><div><div dir="ltr">Hi All,<div><br></div><div>APFloat::fusedMultiplyAdd currently computes the wrong signed zero when small negative results are truncated back to zero in standard precision. The following snippet handles the signedness in fusedMultiplyAdd:</div><div><br></div><font face="monospace, monospace">/* If two numbers add (exactly) to zero, IEEE 754 decrees it is a<br> positive zero unless rounding to minus infinity, except that <br> adding two like-signed zeroes gives that zero. */</font><div><font face="monospace, monospace">if (category == fcZero && sign != addend.sign)<br> sign = (rounding_mode == rmTowardNegative);</font><div><br></div><div>The test "category == fcZero" tells us that the result was zero after rounding back down to standard precision, but since the addition is carried out in extended precision this doesn't guarantee that the result of the addition was exactly zero (so the comment text may not apply). The attached patch adds a check for underflow during truncation which ensures the correct signedness of the result.</div></div></div></div></blockquote><div><br></div><div><br></div></div></div><div>According to what you describe ("the comment text may not apply"), I feel that the comment might be updated as well.</div><div><br></div><div>Best,</div><div><br></div><div>Mehdi</div></div><br></div></div></blockquote></div><br></div>
</div></blockquote></div><br></div></div></div></div></blockquote></div><br></div>