John Lask wrote:
Well this is interesting. So what you are saying is that if your haskell application requires a peek memory utilisation of (for example) 1GB, after the memory intesive computation has completed and the GC has run (assuming all references have been dropped) the GHC RTS will retain the 1GB allocated to the process. Does this occur on both windows and posix platforms, does anyone know ? If so then this is a real issue. It would be reasonable to expect that the RTS release resources to the OS when not explicitly required.
FWIW, the JVM also fails to release memory resources back to the OS. Given all the problems I've seen that one cause for long-running processes, I'm definitely in support of correcting any behavior like this in the GHC RTS. If the RTS does suffer from this, does anyone on this list know the cause (before moving his thread over to ghc@) ? -- Live well, ~wren