<font size=2 face="sans-serif">Hi David,</font><br><br><font size=2 face="sans-serif">Thank you for your comments.</font><br><br><font size=2 face="sans-serif">> Do you have any metrics on this
latest version?</font><br><br><font size=2 face="sans-serif">Pause time of Young GC (3rd-10th in
evaluation period) in SPECjbb2013 was shorten 5.4% and Critical jOPS (which
highly depends on GC pause time) was improved 9.2%. CPU was POWER8 (8247-22L)
and two cores were enabled. 24GB for mx and 20GB for mn.</font><br><br><font size=2 face="sans-serif">> So far the only justification for
making these changes to the GC code </font><br><font size=2 face="sans-serif">> come from the April discussion
[1] where it was stated simply that:</font><br><font size=2 face="sans-serif">> "We've looked at the proposed
changes and we are pretty sure that the</font><br><font size=2 face="sans-serif">> cmpxchg done during copy_to_survivor_space
in the parallel GC doesn't</font><br><font size=2 face="sans-serif">> require the full fence/acquire
semantics." [Martin & Volker]</font><br><font size=2 face="sans-serif">> </font><br><font size=2 face="sans-serif">> Reading back through all the emails,
including the ones in April, I </font><br><font size=2 face="sans-serif">> _think_ part of the reasoning here
is that we're not doing a CAS that </font><br><font size=2 face="sans-serif">> publishes a new object that was
just created, but that we have </font><br><font size=2 face="sans-serif">> previously created that object
using a full CAS and are now only </font><br><font size=2 face="sans-serif">> updating the markword of another
object with a forwarding pointer. The </font><br><font size=2 face="sans-serif">> second cas would not need full
fence semantics as the other object is </font><br><font size=2 face="sans-serif">> already visible. However I am not
a GC expert and other comments by GC </font><br><font size=2 face="sans-serif">> folk suggest that is not in fact
the case, or at least is not </font><br><font size=2 face="sans-serif">> necessarily always the case. So
I can not establish that what is being </font><br><font size=2 face="sans-serif">> proposed is correct.</font><br><font size=2 face="sans-serif">> </font><br><font size=2 face="sans-serif">> I think the GC experts need to
have a discussion to resolve things to </font><br><font size=2 face="sans-serif">> their mutual satisfaction.</font><br><br><font size=2 face="sans-serif">Thank you for lots of your comments
and suggestions. And lots of my mistakes made the discussion long. very
sorry. I would like to know comments of GC experts.</font><br><br><font size=2 face="sans-serif">Regards,<br>Hiroshi<br>-----------------------<br>Hiroshi Horii, Ph.D.<br>IBM Research - Tokyo</font><tt><font size=2><br></font></tt><BR>