<html><head></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; "><br><div><div>On 08 Feb 2013, at 19:06, Florian Klämpfl wrote:</div><br class="Apple-interchange-newline"><blockquote type="cite"><span class="Apple-style-span" style="border-collapse: separate; font-family: Monaco; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: 2; text-align: -webkit-auto; text-indent: 0px; text-transform: none; white-space: normal; widows: 2; word-spacing: 0px; -webkit-border-horizontal-spacing: 0px; -webkit-border-vertical-spacing: 0px; -webkit-text-decorations-in-effect: none; -webkit-text-size-adjust: auto; -webkit-text-stroke-width: 0px; font-size: medium; ">Am 08.02.2013 16:02, schrieb Jonas Maebe:<br><blockquote type="cite"><br></blockquote><blockquote type="cite">On 07 Feb 2013, at 16:52, Ewald wrote:<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite"><blockquote type="cite">Altough I still<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">don't see how you can make these calls atomic then (without the<br></blockquote></blockquote><blockquote type="cite"><blockquote type="cite">barrier)?<br></blockquote></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite">On x86, it is impossible to perform an atomic access without a (partial)<br></blockquote><blockquote type="cite">memory barrier that affects all memory (due to the behaviour of the<br></blockquote><blockquote type="cite">"lock" prefix). On other architectures, the instructions for atomic<br></blockquote><blockquote type="cite">accesses may only lock the cacheline or memory page containing the value<br></blockquote><blockquote type="cite">to be atomically modified. I.e., they are only a barrier for that<br></blockquote><blockquote type="cite">cacheline/page, not for the entire memory.<br></blockquote><br>This might change with Haswell though (when the new instructions are used).<br></span></blockquote></div><br><div>It is already not a memory barrier for certain SSE instructions, that's why I wrote "(partial)".</div><div><br></div><div><br></div><div>Jonas</div></body></html>