This is an internship project to analyze, document, publish and improve our cycle collector. We would like to gain a better understanding of our current implementation, in particular those parts that go beyond the original Bacon paper (GC heaps, separate GC and CC, mark bitmaps, etc). Some measurements about efficiency would be nice, and tools to better understand the CC model, what cycles we find, how long they last, etc. These results should be publishable, and maybe we come up with improvements to the code as well (in particular scheduling heuristics and the data structure representing the CC heap model). Start: February 28 Duration: 4 month
Correction. We can start on the 21st. We have an intern for this. I will add the name when the paperwork is done.
You need to log in before you can comment on or make changes to this bug.