How to read the output of +PrintTenuringDistribution

Tao Mao tao.mao at oracle.com
Thu Oct 24 23:40:40 PDT 2013


Hi Ji,

 From what you've reported, it definitely looks weird. Are these two 
GC's consecutive two GC's (i.e. no other GC/s in between)?

Thanks.
Tao

On 10/24/13 9:45 PM, Ji Cheng wrote:
> Hello,
>
> I have gc log enabled with -XX:+PrintTenuringDistribution. But I'm 
> quite confused with the tenuring distribution below.
>
> =============
> 2013-10-19T19:46:30.244+0800: 169797.045: 
> [GC2013-10-19T19:46:30.244+0800: 169797.045: [ParNew
> Desired survivor size 87359488 bytes, new threshold 4 (max 4)
> - age   1:   10532656 bytes,   10532656 total
> - age   2:   14082976 bytes,   24615632 total
> - age   3:   15155296 bytes,   39770928 total
> - age   4:   13938272 bytes,   53709200 total
> : 758515K->76697K(853376K), 0.0748620 secs] 
> 4693076K->4021899K(6120832K), 0.0756370 secs] [Times: user=0.42 
> sys=0.00, real=0.07 secs]
> 2013-10-19T19:47:10.909+0800: 169837.710: 
> [GC2013-10-19T19:47:10.909+0800: 169837.711: [ParNew
> Desired survivor size 87359488 bytes, new threshold 4 (max 4)
> - age   1:    9167144 bytes,    9167144 total
> - age   2:    9178824 bytes,   18345968 total
> - age   3:   16101552 bytes,   34447520 total
> - age   4:   21369776 bytes,   55817296 total
> : 759449K->63442K(853376K), 0.0776450 secs] 
> 4704651K->4020310K(6120832K), 0.0783500 secs] [Times: user=0.43 
> sys=0.00, real=0.07 secs]
> =============
>
> From What I read, there are 10532656 bytes in age 1 (survived from 1 
> GC) in the first gc. In the second gc, 9178824 bytes in age 2 
> (survived from 2 GCs). This is fine since some objects died between 
> the first and second GC.
>
> But in the second GC, 16101552 bytes are in age 3 while only 14082976 
> bytes in age 2 in the first GC. I don't why this number is increasing. 
> Shouldn't all bytes in age n come from age n-1 in the previous GC? Or 
> I misinterpreted those numbers?
>
> btw, the jvm version is 1.7.0_40.
>
> Thanks.
>
> Ji Cheng
>
>
> _______________________________________________
> hotspot-gc-use mailing list
> hotspot-gc-use at openjdk.java.net
> http://mail.openjdk.java.net/mailman/listinfo/hotspot-gc-use
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.openjdk.java.net/pipermail/hotspot-gc-use/attachments/20131024/33cc2890/attachment.html 


More information about the hotspot-gc-use mailing list