<html><body><p><font size="2">Dear all</font><font size="2">,</font><br><br><font size="2">I would like to ask reviews on 8154736 </font><font size="2">“</font><font size="2">enhancement of cmpxchg and copy_to_survivor</font><font size="2">”</font><font size="2">. The change adds options to avoid expensive syncs with compare-and-exchange. An experiment by using SPECjbb2015 showed 6% improvement in critical-jOPS. This change focuses on ppc64 but would be potentially beneficial for aarch64.</font><br><br><font size="2">Although discussions stopped at </font><a href="http://mail.openjdk.java.net/pipermail/ppc-aix-port-dev/2016-October/002718.html"><font size="2">http://mail.openjdk.java.net/pipermail/ppc-aix-port-dev/2016-October/002718.html</font></a><font size="2">,</font><font size="2"> I would like to restart the review by taking over Hiroshi's work if the discussion is still open.</font><br><br><font size="2">Bug: </font><a href="https://bugs.openjdk.java.net/browse/JDK-8154736"><font size="2">https://bugs.openjdk.java.net/browse/JDK-8154736</font></a><br><font size="2">Webrev: </font><a href="http://cr.openjdk.java.net/~mhorie/8154736/webrev.08/"><font size="2">http://cr.openjdk.java.net/~mhorie/8154736/webrev.08/</font></a><br><br><font size="2">Previous review had discussions on improper log output </font><font size="2">(</font><a href="http://mail.openjdk.java.net/pipermail/ppc-aix-port-dev/2016-September/002672.html"><font size="2">http://mail.openjdk.java.net/pipermail/ppc-aix-port-dev/2016-September/002672.html</font></a><font size="2">)</font><font size="2">. Logs can be generated properly with this change</font><font size="2">,</font><font size="2"> but I would like to ask if we should use </font><font size="2">“</font><font size="2">if</font><font size="2">(</font><font size="2">log</font><font size="2">)</font><font size="2"> OrderAccess:acquire</font><font size="2">()”</font><font size="2"> as is in webrev or more general approach with a call to OrderAccess:consume</font><font size="2">()</font><font size="2"> with empty implementation on all supported platforms.</font><br><br><font size="2">Also</font><font size="2">,</font><font size="2"> there were discussions on the problem of unawareness of copied obj </font><font size="2">(</font><a href="http://mail.openjdk.java.net/pipermail/ppc-aix-port-dev/2016-October/002696.html"><font size="2">http://mail.openjdk.java.net/pipermail/ppc-aix-port-dev/2016-October/002696.html</font></a><font size="2">)</font><font size="2">. This change adds </font><font size="2">“</font><font size="2">release</font><font size="2">”</font><font size="2"> in cmpxchg_pre_membar. This was discussed in </font><a href="http://mail.openjdk.java.net/pipermail/ppc-aix-port-dev/2016-October/002698.html"><font size="2">http://mail.openjdk.java.net/pipermail/ppc-aix-port-dev/2016-October/002698.html</font></a><font size="2">.</font><br><br><font size="2">I measured SPECjbb2015 with its multi JVMs mode on a POWER8 node </font><font size="2">(</font><font size="2">for JDK11</font><font size="2">,</font><font size="2"> I modified MANIFEST in specjbb2015.jar to specify locations of JAXB related libraries</font><font size="2">)</font><font size="2">. As a result</font><font size="2">,</font><font size="2"> critical-jOPS improved by 6% due to this change.</font><br><br><font size="2">Best regards,</font><br><font size="2">--</font><br><font size="2">Michihiro,</font><br><font size="2">IBM Research - Tokyo</font><BR>
</body></html>