JEP 271: Unified GC Logging

[Kirk Pepperdine]

> On Sep 25, 2015, at 10:50 AM, Bengt Rutisson <bengt.rutisson at oracle.com> wrote:
> 
> 
> Hi Kirk,
> 
> On 2015-09-25 09:30, Kirk Pepperdine wrote:
>> Hi Benqt,
>> 
>> I have one question. Is there a reason that the logging is reported in units of 0.1M instead of K. I ask only because I sometimes find that resolution leads to some significant rounding errors in my calculations. Can we move to have it report either in K or just plain bytes.
> 
> I think the reason is that it is nice to have a common unit and it was probably thought that megabytes was enough for many use cases.

Sounds reasonable but the unit of measure is very close to the common region size and having units of measure close to the size of the thing you are measuring is generally never a good idea. It’s like using a ms timer to measure an event that lasts about 1ms. Sometimes you get 2ms, sometimes you get 1ms and sometimes you’ll get 0ms. It all depends on how you are aligned with the clock edge. But 1ms +/- 1ms is a pretty large error. ;-)

> The different GCs use different approaches for the memory units. I currently don't have any plans on fixing that as part of my initial pass on the GC logging. But that is a good thing to look at as a second step. I'd like to get existing logging converted and then fine tune it.

If it’s really a big change than it’s understandable.

Kind regards,
Kirk

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 496 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <https://mail.openjdk.org/pipermail/hotspot-gc-dev/attachments/20150925/90d1d14b/signature.asc>