Comment #27 on issue 2741 by [email protected]: max_opt_count is
implemented using the wrong kind of limit
http://code.google.com/p/v8/issues/detail?id=2741
Yes I've digressed into a tangent on my quest to try and depreciate the
need for the max_opt_count value. This is fun, I'll continue my testing to
see if I can ever succeed and if so I'll try to make an issue report far
more clear and with a proposed patch which hopefully isn't too terrible. :)
Just my final words to bid you a nice night, bleeding_edge has definitely
improved, so it got a little hard to see the effect I was trying to
understand, but awesome none-the-less!
That weird addition to clear the type feedback unoptimized->optimizable()
oddly did change runtime performance, I don't want to push it or anything,
but I'll try to figure out why and report back if I find out what I saw.
And the testing for the state in the runtime-profiler.cc somehow stabilized
the runs of CodeLoad among multiple runs without a full restart of the
browser, it stays more consistently high at around 16,800, rather than
regressing to 15,200 when not being tested.
Well, it's been fun and I'll keep messing around, and I'll chime in if I
actually figure something out if you and the team don't already do it in
some roundabout amazing and correct way as a side-effect from just normal
development of the VM.
Have a good one and farewell for now!
--
You received this message because this project is configured to send all
issue notifications to this address.
You may adjust your notification preferences at:
https://code.google.com/hosting/settings
--
--
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev
---
You received this message because you are subscribed to the Google Groups "v8-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.
