[Rev 02] RFR: 6643: Make VM Operations rule consider close consecutive events in score

Henrik Dafgård hdafgard at openjdk.java.net
Wed Feb 26 13:35:37 UTC 2020


On Mon, 24 Feb 2020 17:13:24 GMT, Jessye Coleman-Shapiro <jescolem at openjdk.org> wrote:

>> The only guarantee is that the lanes are sorted on event times and are disjoint. So that might be a possible outcome.
> 
> Ok,  I have left the sorting of events in my most recent commit https://github.com/openjdk/jmc/pull/42/commits/19fb291dbc838aeb583834690cce663f3a114f91 to deal with the case mentioned above. Although, yes, it is expensive and unnecessary to sort the events if there is only one lane, if there are multiple lanes, this sorting needs to occur to ensure that the events are being compared in the order the occur in real time.  
> 
> Let me know what you think.

I think this might be alright as long as the amount of VM Operation events are fairly low (< 1000) but if there are too many then this approach would be quite slow. At that point we could file a bug to improve the performance here.

-------------

PR: https://git.openjdk.java.net/jmc/pull/42


More information about the jmc-dev mailing list