Garbage Collection is Fast, But a Stack is Faster
Prompted by claims that garbage collection can outperform stack allocation when sufficient physical memory is available, we present a careful analysis and set of cross-architecture measurements comparing these two approaches for the implementation of continuation (procedure call) frames. When the fr...
| Main Authors: | , |
|---|---|
| Language: | en_US |
| Published: |
2004
|
| Online Access: | http://hdl.handle.net/1721.1/6622 |
| _version_ | 1826201251268788224 |
|---|---|
| author | Miller, James S. Rozas, Guillermo J. |
| author_facet | Miller, James S. Rozas, Guillermo J. |
| author_sort | Miller, James S. |
| collection | MIT |
| description | Prompted by claims that garbage collection can outperform stack allocation when sufficient physical memory is available, we present a careful analysis and set of cross-architecture measurements comparing these two approaches for the implementation of continuation (procedure call) frames. When the frames are allocated on a heap they require additional space, increase the amount of data transferred between memory and registers, and, on current architectures, require more instructions. We find that stack allocation of continuation frames outperforms heap allocation in some cases by almost a factor of three. Thus, stacks remain an important implementation technique for procedure calls, even in the presence of an efficient, compacting garbage collector and large amounts of memory. |
| first_indexed | 2024-09-23T11:48:40Z |
| id | mit-1721.1/6622 |
| institution | Massachusetts Institute of Technology |
| language | en_US |
| last_indexed | 2024-09-23T11:48:40Z |
| publishDate | 2004 |
| record_format | dspace |
| spelling | mit-1721.1/66222019-04-11T02:52:34Z Garbage Collection is Fast, But a Stack is Faster Miller, James S. Rozas, Guillermo J. Prompted by claims that garbage collection can outperform stack allocation when sufficient physical memory is available, we present a careful analysis and set of cross-architecture measurements comparing these two approaches for the implementation of continuation (procedure call) frames. When the frames are allocated on a heap they require additional space, increase the amount of data transferred between memory and registers, and, on current architectures, require more instructions. We find that stack allocation of continuation frames outperforms heap allocation in some cases by almost a factor of three. Thus, stacks remain an important implementation technique for procedure calls, even in the presence of an efficient, compacting garbage collector and large amounts of memory. 2004-10-08T20:34:37Z 2004-10-08T20:34:37Z 1994-03-01 AIM-1462 http://hdl.handle.net/1721.1/6622 en_US AIM-1462 94049 bytes 389294 bytes application/octet-stream application/pdf application/octet-stream application/pdf |
| spellingShingle | Miller, James S. Rozas, Guillermo J. Garbage Collection is Fast, But a Stack is Faster |
| title | Garbage Collection is Fast, But a Stack is Faster |
| title_full | Garbage Collection is Fast, But a Stack is Faster |
| title_fullStr | Garbage Collection is Fast, But a Stack is Faster |
| title_full_unstemmed | Garbage Collection is Fast, But a Stack is Faster |
| title_short | Garbage Collection is Fast, But a Stack is Faster |
| title_sort | garbage collection is fast but a stack is faster |
| url | http://hdl.handle.net/1721.1/6622 |
| work_keys_str_mv | AT millerjamess garbagecollectionisfastbutastackisfaster AT rozasguillermoj garbagecollectionisfastbutastackisfaster |