[Gc] Re: Understanding why GC'ing increases to the double in time?

Martin Egholm Nielsen martin at egholm-nielsen.dk
Thu Mar 9 02:22:48 PST 2006


Boehm -> ">>"
Egholm -> ">"

>> Building the collector without SILENT defined will get you more 
>> statistics, which might be useful.  In particular, it should be
>> possible to verify that the amount of live heap memory is not
>> drastically increasing.
> I'll do that...
And now, a month later, here goes. Attached is two log files, with 
succeeding gc-dump information - one with short worldstop-time and 
another with the long worldstop-time. Below is a summary of these files, 
but also scanning the logs does not make me able to see any big differences.

>> The collector is almost certainly spending it's time in mark_from.
>> It would be mildly useful to confirm that. It would be more useful
>> to confirm that this is not being called through the finalization
>> code.
I reckon this still applies - I'll take a look into GC_mark_from in a 
minute and see if I can measure the time spent in there...

LOGS SUMMARY:
=============

=== From short ===
==================

Initiating full world-stop collection 21 after 2784036 allocd bytes
0 bytes in heap blacklisted for interior pointers
--> Marking for collection 21 after 2784036 allocd bytes + 19052 wasted 
bytes
Collection 20 reclaimed 2750772 bytes ---> heapsize = 6088704 bytes
World-stopped marking took 230 msecs
---- 8< 8< 8< ----
blocks = 4605, bytes = 5123072

***Finalization statistics:
5021 finalization table entries; 12 disappearing links
0 objects are eligible for immediate finalization
Bytes recovered before sweep - f.l. count = -15344
Immediately reclaimed 2237448 bytes in heap of size 6088704 bytes
1007932 (atomic) + 806484 (composite) collectable bytes in use
Finalize + initiate sweep took 660 + 20 msecs
Complete collection took 920 msecs

=== From long ===
=================

Initiating full world-stop collection 22 after 3314376 allocd bytes
0 bytes in heap blacklisted for interior pointers
--> Marking for collection 22 after 3314376 allocd bytes + 27184 wasted 
bytes
Collection 21 reclaimed 2843464 bytes ---> heapsize = 6088704 bytes
World-stopped marking took 430 msecs
---- 8< 8< 8< ----
blocks = 4374, bytes = 5645312

***Finalization statistics:
5237 finalization table entries; 5 disappearing links
0 objects are eligible for immediate finalization
Bytes recovered before sweep - f.l. count = -13880
Immediately reclaimed 2105796 bytes in heap of size 6088704 bytes
1024056 (atomic) + 1430164 (composite) collectable bytes in use
Finalize + initiate sweep took 690 + 20 msecs
Complete collection took 1160 msecs

-------------- next part --------------
A non-text attachment was scrubbed...
Name: logs-2.zip
Type: application/zip
Size: 17166 bytes
Desc: not available
Url : http://napali.hpl.hp.com/pipermail/gc/attachments/20060309/bfca60dd/logs-2.zip


More information about the Gc mailing list