I’ve read a couple of quite interesting papers this week.
There’s an interesting paper on memory models with some associated discussion on Lambda The Ultimate. Current memory models are either badly underspecified or are hard to relate to a user’s code, and many models appear to have undesirable side-effects. The paper discusses some of these issues, and potential ways forward.
Cliff Click’s blog has a link to some slides from one of his presentations which discusses Azul System’s aims via its Managed Runtime Initiative to get better support inside Linux for their high throughput garbage collector.