<html><head><meta http-equiv="Content-Type" content="text/html; charset=utf-8"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;" class="">I see nontemporals as a nice builtin so you don’t have to hand-write the assembly. I don’t think existing hardware is consistent enough to allow us to expose nice semantics at a language or IR level, in the sense that we probably don’t want to define what happens if you step out of the comfortable uses for nontemporals (i.e. what happens when you start having temporality). A builtin pre / post fence seems fine to me, but that still leaves a gaping semantic hole… which I’m also fine with.<div class=""><br class=""></div><div class="">FWIW you have the same problem, somewhat worse, with some vector load / store instructions on some ISAs.</div><div class=""><br class=""><div><br class=""><blockquote type="cite" class=""><div class="">On Apr 28, 2020, at 2:42 PM, Cranmer, Joshua <<a href="mailto:joshua.cranmer@intel.com" class="">joshua.cranmer@intel.com</a>> wrote:</div><br class="Apple-interchange-newline"><div class=""><div class="WordSection1" style="page: WordSection1; caret-color: rgb(0, 0, 0); font-family: Helvetica; font-size: 12px; font-style: normal; font-variant-caps: normal; font-weight: normal; letter-spacing: normal; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration: none;"><div style="margin: 0in 0in 0.0001pt; font-size: 11pt; font-family: Calibri, sans-serif;" class="">The current specification of the behavior of the !nontemporal attribute in LLVM, and the __builtin_nontemporal_* functions in Clang, is rather spartan and underspecified. In effect, it says the following things:<o:p class=""></o:p></div><ul type="disc" style="margin-bottom: 0in; margin-top: 0in;" class=""><li class="MsoListParagraph" style="margin: 0in 0in 0.0001pt; font-size: 11pt; font-family: Calibri, sans-serif;">Atomic !nontemporal has no defined semantics<o:p class=""></o:p></li><li class="MsoListParagraph" style="margin: 0in 0in 0.0001pt; font-size: 11pt; font-family: Calibri, sans-serif;">!nontemporal may use special instructions to save cache bandwidth, such as “MOVNT” on x86.<o:p class=""></o:p></li></ul><div style="margin: 0in 0in 0.0001pt; font-size: 11pt; font-family: Calibri, sans-serif;" class=""><o:p class=""> </o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 11pt; font-family: Calibri, sans-serif;" class="">What is crucially lacking from this specification is its effects in relation to other memory ordering constructs, namely fences. In lieu of text to the contrary, one could reasonably come to the conclusion that this code should be sufficient to generate correctly working code:<o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 11pt; font-family: Calibri, sans-serif;" class=""><o:p class=""> </o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 11pt; font-family: Calibri, sans-serif;" class="">store i64 %x, i64* %addr, !nontemporal !100<o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 11pt; font-family: Calibri, sans-serif;" class="">fence release<o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 11pt; font-family: Calibri, sans-serif;" class=""><o:p class=""> </o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 11pt; font-family: Calibri, sans-serif;" class="">But on x86 today, it does not: the first store is lowered to a MOVNT, while the fence release is lowered to a compiler-only memory barrier instead of an SFENCE (as regular load/store/atomicrmw operations already have these operations built into their execution semantics). So the fence instructions appear to be lowered with the expectation that nontemporal instructions do not exist, which is not implied by the language reference.<o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 11pt; font-family: Calibri, sans-serif;" class=""><o:p class=""> </o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 11pt; font-family: Calibri, sans-serif;" class="">I’ve been pointed to an earlier thread (<a href="https://lists.llvm.org/pipermail/llvm-dev/2016-January/093912.html" style="color: rgb(149, 79, 114); text-decoration: underline;" class="">https://lists.llvm.org/pipermail/llvm-dev/2016-January/093912.html</a>) that suggests an introduction of nontemporal-specific fences, but there appears to have been no effort to actually push this into LLVM. What’s the current status of this work, and what is the right way to move forward and clarifying the semantics of !nontemporal with respect to threads.<o:p class=""></o:p></div><div style="margin: 0in 0in 0.0001pt; font-size: 11pt; font-family: Calibri, sans-serif;" class=""><span style="font-size: 9pt;" class=""><o:p class=""> </o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 11pt; font-family: Calibri, sans-serif;" class=""><span style="font-size: 9pt;" class="">--<span class="Apple-converted-space"> </span><o:p class=""></o:p></span></div><div style="margin: 0in 0in 0.0001pt; font-size: 11pt; font-family: Calibri, sans-serif;" class=""><i class=""><span style="font-size: 9pt;" class="">Joshua Cranmer</span></i></div></div></div></blockquote></div><br class=""></div></body></html>