The post showed up as empty (only title) in the RSS feed. It would have been nice if you had posted more information (e.g. the abstract) or a simple sentence as in the previous post (which all showed up correctly for some reason).
Thanks - I'll include the abstract in the future.
Thanks for posting. I am looking at this from the perspective of a Java application developer. The risk of overthreading has become significantly higher with Java 8, because everyone is just throwing parallel()-statements into their stream processing without much thought.
Is your work anything to do with addressing such concerns within the Oracle JVM? For Java 9, say?
Hi Sebastian, I actually did some of the initial prototyping of "concurrency restriction" ideas in hotspot with the synchronized construct. I've not given any more thought, however, as to when, how or if that'd be transferred to the product group. It's possible that Doug can use the idea in JUC -- I've discussed this with him previously.
I agree completely that the risk of overthreading is increasing as we add parallelism constructs such as thread pools, stream parallelism, etc. Somewhat perversely, 5 years ago we were worried about having enough threads to utilize our new multicore systems, now we're starting to worry about having too many.
Malthusian Locks appears in EuroSys 2017. An extended version is in
arxiv.Abstract:Applications running in modern multithreaded...
Pointers: Experiences with HTM-Based Reference Counting in C++by Maria Carpen-Amarie , Dave Dice, Gaël Thomas and Pascal...
The following is inspired by some discussions with John Rose on how we might
implement atomic access to value type "tuples" for subsequent...