Posts

Showing posts with the label emery berger

On "Quantifying the Performance of Garbage Collection vs. Explicit Memory Management"

The computer science research paper " Quantifying the Performance of Garbage Collection vs. Explicit Memory Management " by Emery Berger and Matthew Hertz contains an interesting study about memory management. However, the conclusions given in the paper were badly worded and are now being used to justify an anti-GC ideology. Introduction That paper describes an experiment that analyzed the performance of a benchmark suite using: Tracing garbage collection. Oracular memory management (precomputing the earliest point free could have been inserted). The experiment was performed: On one VM, the Jikes Research Virtual Machine (RVM). Using one programming language, Java, and consequently one programming paradigm, object oriented programming. Using each of the five different garbage collection algorithms provided by that VM. The five GC algorithms are: GenMS - Appel-style generational collector  (1988) GenCopy -  two generations with copying mature space CopyMS -  nursery...