Weak memory has been an active research area for well over a decade now, with many key results published here at POPL. Researchers have built many formal models of how weak memory works (or should work) in a variety of languages and architectures, and they have designed and implemented many analyses that take weak-memory effects into account.
The aim of this workshop is to bring together experts in weak memory from across industry and academia to discuss what we can expect from the next decade of weak memory research. Can we expect more languages and architectures to develop increasingly complicated weak memory models, or will we see a regression to sequential consistency semantics? How might new technologies like persistent memory and CXL complicate the picture? Will AI chatbots obviate the need for human programmers to grapple with which reads can be reordered with which writes?
List of speakers
Here is a very tentative list of speakers, together with some even more tentative titles, and all in no particular order.
Title | |
---|---|
A case against semantic dependencies The Future of Weak Memory | |
Compilers should get over themselves and respect semantic dependencies! The Future of Weak Memory | |
Evolving Weak Memory Models for Evolving Architectures The Future of Weak Memory | |
In-order execution nails every weak memory behavior The Future of Weak Memory | |
Programmers love mind-bogglingly complicated weak memory models The Future of Weak Memory | |
Some things I wish I hadn’t seen The Future of Weak Memory | |
TBC The Future of Weak Memory | |
TBC The Future of Weak Memory | |
TBC The Future of Weak Memory | |
TBC The Future of Weak Memory | |
TBC The Future of Weak Memory | |
TBC The Future of Weak Memory | |
TBC The Future of Weak Memory | |
The Future of Compiler Testing with Weak Memory Models is Relaxed The Future of Weak Memory |
Call for Papers
-
The format of the workshop will be a series of 20-minute informal talks intended to set out a (possibly controversial) position.
-
Talks about early ideas and work-in-progress are welcome.
-
Attendees will be invited (but not obliged) to submit 1-page position papers beforehand.
-
There will be ample time set aside for discussions among participants, and depending on participation levels, we hope to organise a couple of discussion panels.
-
No formal proceedings are planned, but we will offer authors the option of making short position papers available from our workshop’s webpage.
If you’d like to contribute a talk and/or a position paper, please email j.wickerson@imperial.ac.uk.
List of participants/speakers
Here are a few more people who might be participating in the workshop (possibly speaking, or possibly just in the audience):
- Hans Boehm (Google)
- Alan Jeffrey (Roblox)
- Hernán Ponce de Leon (Huawei)
- James Riely (DePaul)
- Derek Dreyer (MPI-SWS)
- Dan Lustig (NVIDIA)
- Luc Maranget (INRIA)
- Anton Podkopaev (JetBrains)
- Conrad Watt (U Cambridge)