JDK-8148960: Humongous mis-spelled in log output

kirk.pepperdine at gmail.com kirk.pepperdine at gmail.com
Wed Feb 3 16:57:24 UTC 2016


Hi Charlie,

There are about 60 flags that drop data into/alter the format of the GC logs. I think (I’ve not tested for all combinations of them) I can cope with the output from about 40ish of them and 1 of the 40 is the output from G1SummarizeRSetStats, I just don’t do much with the data at the moment. A lot of the data is interesting but only for debugging the collectors and it rarely offers little to no value when tuning so I don’t really do anything with it. The huge value is in occupancy before and after along with pause times. Calculations on those numbers alone can give you insight into where to go to get (sometimes huge) gains in performance. If there is anything you’re interested in looking at in a particular way, please just send me a request, I’ll happily drop it in for you.

Regards,
Kirk

> On Feb 3, 2016, at 5:45 PM, charlie hunt <charlie.hunt at oracle.com> wrote:
> 
> I was hoping you didn’t have parsers for G1SummarizeRSetStats. ;-)  And, hopefully you (and others) don’t have to go to that level of detail in your analysis tools.
> 
> charlie
> 
>> On Feb 3, 2016, at 10:32 AM, kirk.pepperdine at gmail.com wrote:
>> 
>> Hi Charlie,
>> 
>> This has been around for so long I was hoping no one would notice it.  Now I’ve got to look for both the correct and incorrect spellings
>> 
>> Reminds me of when Cray wanted to upgrade the floating point multiply unit in their ALU to the IEEE standard. IEEE was far better than what they were using. The problem was; there was so much software that worked around the well known problem and so much reference data accumulated using the “bad” calculation that no one wanted the upgrade as it would have been too disruptive.
>> 
>> Regards,
>> Kirk
>> 
>>> On Feb 3, 2016, at 5:21 PM, charlie hunt <charlie.hunt at oracle.com> wrote:
>>> 
>>> Looks good … though I am not an official Reviewer.
>>> 
>>> If only all bugs could be so trivial to fix. ;-)
>>> 
>>> charlie
>>> 
>>>> On Feb 3, 2016, at 10:18 AM, Bengt Rutisson <bengt.rutisson at oracle.com> wrote:
>>>> 
>>>> 
>>>> Hi all,
>>>> 
>>>> Could I have a couple of reviews for this small change?
>>>> 
>>>> http://cr.openjdk.java.net/~brutisso/8148960/webrev.00/
>>>> https://bugs.openjdk.java.net/browse/JDK-8148960
>>>> 
>>>> Thanks Charlie for finding it. Here's his summary:
>>>> 
>>>> I noticed we have a mis-spelling of humongous in:
>>>> ./vm/gc/g1/g1RemSetSummary.cpp
>>>> RegionTypeCounter _humonguous;
>>>> HRRSStatsIter() : _all("All"), _young("Young"), _humonguous("Humonguous"),
>>>>   current = &_humonguous;
>>>> RegionTypeCounter* counters[] = { &_young, &_humonguous, &_free, &_old, NULL };
>>>> 
>>>> Also note the variable name also happens to be misspelled too. But, that's probably not as as annoying as when you see the mis-spelling in G1 log output.
>>>> 
>>>> This mis-spelling of humongous shows in summarized RSet out produced by G1SummarizeRSetStats.
>>>> 
>>>> Thanks,
>>>> Bengt
>>> 
>> 
> 



More information about the hotspot-gc-dev mailing list