On Fri, 27 Mar 2026 08:49:49 GMT, Alan Bateman <[email protected]> wrote:

> > Do we think it is worth proceeding with this change?
> 
> Thanks for the further context (and to Eirik Bjorsnos for digging into the 
> Cassandra code). I don't know how common it is for pooled threads to change 
> their thread names for the task they are executing but there are at least 
> some cases. For non-trivial tasks then I assume the cost of changing the 
> thread name is in the noise but for tiny tasks then maybe it's an issue. 
> Going forward then I hope there will be less pooling/sharing with tasks 
> executed by virtual threads.
> 
> So I think I agree it's worth trying to improve although I'm a bit dubious 
> about have Thread.setName check if the name is the same as the old name, does 
> that really help?

Well it obviously helps if the name doesn't change. However, that is not the 
usage pattern that would affect Cassandra. According to my measurements, the 
same-name check adds a few ns to the common case. I am not sure if it's useful 
for any real workload, though. I have removed that check.

-------------

PR Comment: https://git.openjdk.org/jdk/pull/30374#issuecomment-4153256415

Reply via email to