RFR (XS) 8156803: Turn StressLCM/StressGCM flags to diagnostic

Aleksey Shipilev aleksey.shipilev at oracle.com
Wed May 11 21:33:00 UTC 2016


In order to better test the concurrency support in JVM, we introduced
StressLCM/StressGCM flags that randomize C2's instruction scheduling:

  develop(bool, StressLCM, false, \
          "Randomize instruction scheduling in LCM") \

  develop(bool, StressGCM, false, \
          "Randomize instruction scheduling in GCM") \

These, however, require fastdebug build, which is not really fast for
concurrency testing, as it does lots of asserts in VM codepaths
(notably, GC). We would be better off turning these flags to diagnostic,
to pop out in product builds.



Testing: manual testing with local x86_64/release build. I don't think
we need to waste machine time on deeply testing this trivial change?


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: OpenPGP digital signature
URL: <http://mail.openjdk.java.net/pipermail/hotspot-compiler-dev/attachments/20160512/401b112a/signature.asc>

More information about the hotspot-compiler-dev mailing list