Hacker News Reader

Comment by pcwalton | Hacker News Reader

> Seems like the main advantage of this transaction oriented collector over a typical generational collector is that it precisely identifies the old-to-young pointers and thus may scan less total objects.

I'm confused as to what this means. Write barriers, just as suggested here, are the standard solution for this issue in generational garbage collectors, and they precisely identify the old-to-young pointers as well. (In fact, the nursery and tenured generations are often separated in memory, so you can test just by masking a few bits off.)

I agree with the sibling comment that this seems to be a simple generational GC, just with a different policy for tenuring than the typical textbook one.

View on HN

The standard solution is that the write barrier updates a card table, which generally reserves a bit for every page. When it comes time to mark through old-to-young pointers, the collector scans the card table - if it sees a bit set in the card table, it must mark through /all objects/ on that page - even if only one of the objects on the page has a cross-generational pointer.

Well, another solution (which I believe is common) is to just tenure the young object (and recursively tenure objects reachable from it) if the write barrier trips, which preserves the precision.

There isn't anything fundamental to generational GC that requires imprecision. It's an implementation detail. In fact, David Ungar's original generational GC paper from 1984 [1] uses precise remembered sets. GHC's GC [2] among others uses precise remembered sets.

[1]: https://people.cs.umass.edu/~emery/classes/cmpsci691s-fall20...

[2]: https://ghc.haskell.org/trac/ghc/wiki/Commentary/Rts/Storage...

Tenuring the young object in the write-barrier is not correct since the the young objects reachable from the newly-tenured object won't be marked in the young collection. That is why TOC then traces this newly-tenured object, which is novel to my knowledge.

Yeah, of course you need to recursively tenure objects, and I should have been more precise about that. I've heard of doing this before; it isn't novel.

My point was to compare TOC to "typical generational collectors" (i.e. those using card-marking, which is common).

My point was not that TOC is fundamentally different from generational collection.

Indeed, as you say, write barriers precisely identify old-to-young pointers at the time of the write. But most collectors defer tracing and track this old-to-young pointer in a card-table, which results in lost precision i.e. now the collector only knows that some object in this region has an old-to-young pointer.

Reply on news.ycombinator.com