<html><head><meta http-equiv="Content-Type" content="text/html charset=utf-8"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" class=""><br class=""><div><blockquote type="cite" class=""><div class="">On Apr 11, 2016, at 12:55 PM, via llvm-dev <<a href="mailto:llvm-dev@lists.llvm.org" class="">llvm-dev@lists.llvm.org</a>> wrote:</div><br class="Apple-interchange-newline"><div class=""><meta http-equiv="Content-Type" content="text/html charset=utf-8" class=""><div style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" class=""><br class=""><div class=""><blockquote type="cite" class=""><div class="">On Apr 11, 2016, at 4:23 AM, Daniel Sanders <<a href="mailto:Daniel.Sanders@imgtec.com" class="">Daniel.Sanders@imgtec.com</a>> wrote:</div><br class="Apple-interchange-newline"><div class=""><div class="WordSection1" style="page: WordSection1; font-family: Helvetica; font-size: 12px; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; widows: auto; word-spacing: 0px; -webkit-text-stroke-width: 0px;"><div style="margin: 0cm 0cm 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">> I am not entirely sure this is safe. Transforming this to an fsub could change the value stored on platforms that implement negates using arithmetic instead of with bitmath (such as ours)<o:p class=""></o:p></div><div style="margin: 0cm 0cm 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class=""><o:p class=""> </o:p></div><div style="margin: 0cm 0cm 0.0001pt; font-size: 12pt; font-family: 'Times New Roman', serif;" class="">I think it's probably safe for IEEE754-2008 conformant platforms because negation was clarified to be a non-arithmetic bit flip that cannot cause exceptions in that specification.</div></div></div></blockquote></div><div class=""><br class=""></div><div class="">I did some digging into IEEE-754 and it seems like this is actually not even safe on fully conformant IEEE-754-2008 platforms.</div><div class=""><br class=""></div><div class=""><div style="margin: 0px; line-height: normal; font-family: Times; -webkit-text-stroke-color: rgb(0, 0, 0); -webkit-text-stroke-width: initial;" class="">5.5.1 Sign bit operations</div><div style="margin: 0px; line-height: normal; font-family: Times; -webkit-text-stroke-color: rgb(0, 0, 0); -webkit-text-stroke-width: initial;" class="">5.5.1.0 Implementations shall provide the following homogeneous quiet-computational sign bit operations for all supported arithmetic formats; they only affect the sign bit. The operations treat floating-point numbers and NaNs alike, and signal no exception. These operations may propagate non-canonical encodings.</div><div style="margin: 0px; line-height: normal; font-family: Times; -webkit-text-stroke-color: rgb(0, 0, 0); -webkit-text-stroke-width: initial; min-height: 14px;" class=""><br class=""></div><div style="margin: 0px; line-height: normal; font-family: Times; -webkit-text-stroke-color: rgb(0, 0, 0); -webkit-text-stroke-width: initial;" class="">copy(x) copies a floating-point operand x to a destination in the same format, with no change to the sign bit. </div><div style="margin: 0px; line-height: normal; font-family: Times; -webkit-text-stroke-color: rgb(0, 0, 0); -webkit-text-stroke-width: initial;" class="">negate(x) copies a floating-point operand x to a destination in the same format, reversing the sign bit. negate(x) is not the same as subtraction(0, x) (see 6.3).</div><div style="margin: 0px; line-height: normal; font-family: Times; -webkit-text-stroke-color: rgb(0, 0, 0); -webkit-text-stroke-width: initial;" class=""><br class=""></div><div style="margin: 0px; line-height: normal; -webkit-text-stroke-color: rgb(0, 0, 0); -webkit-text-stroke-width: initial;" class="">Note the MAY. fneg is required to flip the top bit even if the input is a NaN. But fneg is not required to maintain the other bits. If the input is a non-canonical NaN, the fneg MAY canonicalize it. In fact, even the ‘copy’ MAY canonicalize it. (it also MAY choose to not canonicalize it)</div><div style="margin: 0px; line-height: normal; -webkit-text-stroke-color: rgb(0, 0, 0); -webkit-text-stroke-width: initial;" class=""><br class=""></div><div style="margin: 0px; line-height: normal; -webkit-text-stroke-color: rgb(0, 0, 0); -webkit-text-stroke-width: initial;" class="">Thus, if the integer being fneg’d is a non-canonical NaN, fneg MAY modify bits other than the top bit.</div><div style="margin: 0px; line-height: normal; -webkit-text-stroke-color: rgb(0, 0, 0); -webkit-text-stroke-width: initial;" class=""><br class=""></div><div style="margin: 0px; line-height: normal; -webkit-text-stroke-color: rgb(0, 0, 0); -webkit-text-stroke-width: initial;" class="">—eschar</div></div></div></div></blockquote><br class=""></div><div>[The language frontend plus runtime plus] LLVM is the "implementation”. LLVM can define fneg to be a pure signbit operation if we choose to do so.</div><div><br class=""></div><div>Also, there are no non-canonical encodings in binary16, binary32, or binary64, which I believe are what is under discussion on that thread.</div><div><br class=""></div><div>– Steve</div></body></html>