RFR: 8061254: SPECjvm2008-XML performance regressions in 9-b33

Claes Redestad claes.redestad at oracle.com
Fri May 15 13:46:35 UTC 2015


On 2015-05-15 15:23, Alan Bateman wrote:
> On 13/05/2015 12:51, Claes Redestad wrote:
>> Hi,
>>
>> 9-b33 introduced a sustained regression in SPECjvm2008
>> xml.transform on a number of our test setups.
>>
>> JDK-8058643 removed the check on value.length > 0, which
>> means repeated calls to "".hashCode() now do a store of the
>> calculated value (0) to the hash field. This has potential to
>> cause excessive cache coherence traffic in xml.transform[1]
> I remember the original change but of course didn't spot that empty 
> String would result in a write of 0. I like Shipilev's First Law.

Yes, this is a fine example of that law. :-)

>
> The patch looks good to me.

Thanks!

/Claes



More information about the core-libs-dev mailing list