Project Loom presentation at BangaloreJUG

[Nawazish Khan]

Thanks Alan, I am curious and want to understand what you mean by pinning
of carrier thread. Seems like an important concept because we even have a
JVM switch for it.

Sincerely,
MNK

On Mon, Dec 16, 2019, 2:05 AM Alan Bateman <Alan.Bateman at oracle.com> wrote:

> On 14/12/2019 11:34, Nawazish Khan wrote:
> > Hi All,
> >
> > I would be speaking on Project Loom at BangaloreJUG which would be hosted
> > at Oracle office on 21st December, 2019.
> >
> > Let me know if I could help by taking any survey or any other information
> > which might be helpful to all of us as an input to the project Loom.
> >
> We don't have any prepared surveys but there are several areas where it
> would be useful to collect data.
>
> In the current prototype, if a virtual thread parks (network I/O or
> locking operation) while holding a monitor then it pins the carrier
> thread. This is a quality-of-service rather than a correctness issue but
> it would be useful to get reaction to needing to change some (not all)
> code to use java.util.concurrent locks to workaround this limitation for
> the short term. If there are developers trying out the current repo in
> the meet-up then get them to run with -Djdk.tracePinnedThreads=full and
> see if the monitors usages would be easy or hard to change.
>
> Thread locals. It would be useful to get data on whether developers are
> using TLs directly or not. We've looked at a lot of TL usages in
> libraries and frameworks but not applications as they aren't easy to get
> at. One of the topics being explored in this project is scope locals and
> it would be useful to get an idea how many of the TL usages could be
> replaced with scope locals.
>
> -Alan.
>
>
>
>