<div class="gmail_quote">On Tue, Nov 29, 2011 at 14:09, Dmitry Karpeev <span dir="ltr"><<a href="mailto:karpeev@mcs.anl.gov">karpeev@mcs.anl.gov</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div class="im"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div>I wouldn't consider that stackoverflow post to be authoritative, but there is a large body of literature on lock-free synchronization. </div>


</blockquote><div><br></div></div><div>Could you point me to some of it?</div></blockquote><div><br></div><div>These are not bad places to start</div><div><br></div><div><a href="https://www.google.com/search?q=lock-free">https://www.google.com/search?q=lock-free</a></div>
<div><a href="http://scholar.google.com/scholar?q=lock-free">http://scholar.google.com/scholar?q=lock-free</a></div><div><br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div class="im"><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div>Unless there is resource contention, spinning is always the lowest latency choice. All multi-threaded cores have some sort of pause instruction because it's necessary to make that design useful. For this purpose, we should just spin.</div>



<div><br></div><div>Note that with sequential consistency, </div></blockquote><div><br></div></div><div>What do you mean by "sequential consistency"? If we did serialize the threads before waking them up?</div>
</blockquote></div><br><div>Sequential consistency has to do with the memory model.</div><div><br></div><div><a href="http://en.wikipedia.org/wiki/Sequential_consistency">http://en.wikipedia.org/wiki/Sequential_consistency</a></div>