RFR (S): 8019902: G1: Use the average heap size rather than the minimum heap size to calculate the region size

Bengt Rutisson bengt.rutisson at oracle.com
Thu Aug 29 05:34:47 PDT 2013


Hi Thomas,

Thanks for looking at this!

I agree with you about the log message. Here is an updated webrev 
without the new logging:

http://cr.openjdk.java.net/~brutisso/8019902/webrev.01/

Thanks,
Bengt


On 8/29/13 2:28 PM, Thomas Schatzl wrote:
> Hi,
>
> On Thu, 2013-08-29 at 11:26 +0200, Bengt Rutisson wrote:
>> Hi everyone,
>>
>> Could I have a couple of reviews of this change:
>>
>> http://cr.openjdk.java.net/~brutisso/8019902/webrev.00/
>>
>> The fact that G1 by default bases its region size on the minimum heap
>> size means that out of the box the region size will always be 1M. This
>> is a problem on large machines with lots of memory. We pick a large heap
>> size but get a very small region size. The small regions are inefficient
>> and cause a lot of memory footprint. Normally we aim to get around 2048
>> regions, but on a machine with a lot of memory we might pick a default
>> max heap size of 32G, which means that we will get ~32000 regions. This
>> can lead to out of memory situations - especially on Solaris x86.
>>
>> This patch changes the heuristics for picking the region size to use the
>> average between initial heap size (-Xms) and the maximum heap size
>> (-Xmx). This means that for large heaps we will pick larger region
>> sizes. In the 32G example we will now pick a region size of 8m which
>> means that we will have 4000 regions which is more reasonable.
>    Looks good.
>
> One minor nit: I do not think that the additional log message in
> heapRegion.cpp adds more information than is already printed by other
> log messages (PrintGCDetails, PrintFlagsFinal, jcmd, and others).
>
> This would also not require the change that initializes the G1Log
> earlier than before.
>
> Thanks,
>    Thomas
>
>



More information about the hotspot-gc-dev mailing list