RFR(XXS): 8027454: Do not traverse string table during G1 remark when treating them as strong roots during initial mark (original) (raw)

Mikael Gerdin mikael.gerdin at oracle.com
Fri Jan 10 17:07:58 UTC 2014


On Friday 10 January 2014 10.34.14 Thomas Schatzl wrote:

Hi all,

could I have some reviews for the following tiny change? It disables string table unlink in the G1 remark pause. This is not required as G1 initial mark treats them as strong roots always. This needs to be changed again as soon as class unloading during concurrent mark is implemented, but for now it gives a performance boost about 10-15% of the remark pause across many applications. It is based on the recently posted RFR for JDK-8027476. Webrev: http://cr.openjdk.java.net/~tschatzl/8027454/webrev/

Looks good, Ship it! /Mikael

bugs.sun.com: https://bugs.openjdk.java.net/browse/JDK-8027454 Testing: jprt, FMW apps, specjbb2005, specjbb2013, dacapo, and lots of other testing in the recent months Thanks, Thomas

Thanks, Thomas



More information about the hotspot-gc-dev mailing list