T. Paul Lee and R. E. Barkley. A watermark-based lazy buddy system for kernel memory allocation. In Proceedings of the 1989.  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... reasonable for the Deallocate time bound to be amortized, because the result of the operation is not important and the actual work can be delayed until the CPU is idle (or the memory becomes full) Indeed, this delay idea has been used to improve the cost of the standard buddy system s Deallocate [1, 10, 15, 16]. On the other hand, for the purposes of theoretical results, it would be of interest to obtain a constant worst case time bound for both Allocate and Deallocate. ....

T. Paul Lee and R. E. Barkley. A watermark-based lazy buddy system for kernel memory allocation. In Proceedings of the 1989 Summer USENIX Conference, pages 1--13, June 1989.

....node size policy uses a single predetermined node size for all nodes CHAPTER 4. GUARDED PAGE TABLE EVALUATION 36 within the GPT tree. This has the advantage of simplicity. It removes the complexity of having to manage multiple sized memory objects which is a low, but not negligible overhead [PT77, LB89] It also avoids any complexity associated with choosing a node size. However, a single node size policy has the disadvantage of not adapting to the memory layout of the application, i.e. it could use more memory for a given tree depth than a scheme using a multiple size policy. Section 4.3.1 ....

T. Paul Lee and R. E. Barkley. A watermark-based lazy buddy system for kernel memory allocation. In Proc Summer USENIX Conf., 1989.

....the first dozen fields of an inode (48 bytes) are frequently referenced. Then the majority of inode related memory traffic will be Such allocators are common because they are easy to implement. For example, 4. 4BSD and SVr4 both employ power of two methods [McKusick88, Lee89]. at addresses between 0 and 47 modulo 512. Thus the cache lines near 512 byte boundaries will be heavily loaded while the rest lie fallow. In effect only 9 (48 512) of the cache will be usable by inodes. Fully associative caches would not suffer this problem, but current hardware trends are ....

....bus 0. The situation is exacerbated by an inflated miss rate, since all of the inodes are fighting over a small fraction of the cache. These effects can be dramatic. On a SPARCcenter 2000 running LADDIS under a SunOS 5. 4 development kernel, replacing the old allocator (a power of two buddy system [Lee89]) with the slab allocator reduced bus imbalance from 43 to just 17 . In addition, the primary cache miss rate dropped by 13 . 4.3. Slab Coloring The slab allocator incorporates a simple coloring scheme that distributes buffers evenly throughout the cache, resulting in excellent cache utilization ....

[Article contains additional citation context not shown here]

T. Paul Lee and R. E. Barkley, A Watermark-based Lazy Buddy System for Kernel Memory Allocation. Proceedings of the Summer 1989 Usenix Conference, pp. 1-13.

No context found.

T. Paul Lee and R. E. Barkley. A watermark-based lazy buddy system for kernel memory allocation. In Proceedings of the 1989.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC