RFR: 8166840: Synthetic bridge constructor in ArrayList$Itr blocks inlining

[Claes Redestad]

Hi Kris,

right, I don't intend to meander away on a micro-optimization spree and 
chase down every case where we're hurt by this corner case in the JDK 
and elsewhere (I did check that other core collection classes aren't 
affected, though ;-)), but to get this simple and possibly high-impact 
quick fix out of the way to unblock other work I'm doing, specifically 
JDK-8161210

I think everyone would be very happy to get the root issue in HotSpot fixed!

Thanks!

/Claes

On 2016-09-28 16:40, Krystal Mok wrote:
> Hi Claes,
>
> For this particular case, this JDK-side change looks good to me.
>
> Let me post out the HotSpot version of the change and let you guys 
> decide whether or not you guys want to take that version (which will 
> take care of the ArrayList$1 case without the JDK-side change).
>
> Thanks,
> Kris
>
> On Wed, Sep 28, 2016 at 4:48 AM, Claes Redestad 
> <claes.redestad at oracle.com <mailto:claes.redestad at oracle.com>> wrote:
>
>     Hi,
>
>     as discussed recently on hotspot-compiler-dev[1], having a private
>     class with no default constructor can lead to C2 failing to
>     inline, due to the synthetic bridge constructor using a dummy
>     argument of an uninitialized class. This is really a problem in
>     C2, as well as something which could ultimately be resolved by
>     nestmates...
>
>     However, there is an easy workaround in adding an empty
>     package-private constructor. In the most recently found case - a
>     microbenchmark stressing MethodHandles.iteratedLoop - adding this
>     to ArrayList$Itr lead to a 2.5-3x speedup.
>
>     This is me asking nicely to do a quick-fix for this in
>     java.util.ArrayList$Itr now:
>
>     Bug: https://bugs.openjdk.java.net/browse/JDK-8166840
>     <https://bugs.openjdk.java.net/browse/JDK-8166840>
>     Webrev: http://cr.openjdk.java.net/~redestad/8166840/webrev.01/
>     <http://cr.openjdk.java.net/%7Eredestad/8166840/webrev.01/>
>
>     Thanks!
>
>     /Claes
>
>     [1]
>     http://mail.openjdk.java.net/pipermail/hotspot-compiler-dev/2016-September/024407.html
>     <http://mail.openjdk.java.net/pipermail/hotspot-compiler-dev/2016-September/024407.html>
>
>