[concurrency-interest] ThreadLocal vs ProcessorLocal

Mark Thornton mthornton at optrak.com
Thu Oct 18 03:36:48 EDT 2012


On 18/10/12 04:46, Jacy Odin Grannis wrote:
> So, when I'd seen this it was on Linux 2.6.9, Opteron 8384s, Java
> 1.6.0_16 (it was a couple years back).
>
> Based on this:
> http://juliusdavies.ca/posix_clocks/clock_realtime_linux_faq.html  it
> would look like it could be that the kernel was buggy, as only 2.6.18
> and up are solid.
>
> However, I also found this link:
> http://efreedom.com/Question/1-6814792/Clock-Gettime-Erratic  which
> notes erratic behavior even on 2.6.26, when using an Opteron.
> Somewhere in this thread
> http://stackoverflow.com/questions/510462/is-system-nanotime-completely-useless
> someone comments that when they used AMD before that there wasn't any
> synchronization even across cores on the same die.
I seem to remember that when AMD first introduced per core control of 
the processor clock, the TSC counter on each core reflected the clock 
rate on that core. It took a while for operating systems to check for 
this behaviour, and subsequently the TSC was changed to tick at a 
constant rate regardless of power management.

Mark



More information about the Concurrency-interest mailing list