Data layouts for object-oriented programs - hirzels.com.pptVIP

下载本文档

5
0
约3.75千字
约 21页
2016-10-31 发布于广东
举报
版权申诉

Data layouts for object-oriented programs - hirzels.com.ppt

1、原创力文档（book118）网站文档一经付费（服务费），不意味着购买了该文档的版权，仅供个人/单位学习、研究之用，不得用于商业用途，未经授权，严禁复制、发行、汇编、翻译或者网络传播等，侵权必究。。
2、本站所有内容均由合作方或网友上传，本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺！文档内容仅供研究参考，付费前请自行鉴别。如您付费，意味着您自己接受本站规则且自行承担风险，本站不退款、不进行额外附加服务；查看《如何避免下载的几个坑》。如果您已付费下载过本站文档，您可以点击这里二次下载。
3、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等，请点击“版权申诉”（推荐），也可以打举报电话：400-050-0827(电话支持时间：9:00-18:30)。
4、该文档为VIP文档，如果想要下载，成为VIP会员后，下载免费。
5、成为VIP后，下载本文档将扣除1次下载权益。下载后，不支持退款、换文档。如有疑问请联系我们。
6、成为VIP后，您将拥有八大权益，权益包括：VIP文档下载权益、阅读免打扰、文档格式转换、高级专利检索、专属身份标志、高级客服、多端互通、版权登记。
7、VIP文档为合作方或网友上传，每下载1次，网站将根据用户上传文档的质量评分、类型等，对文档贡献者给予高额补贴、流量扶持。如果你也想贡献VIP文档。上传文档

Data layouts for object-oriented programs - hirzels.com.ppt

Data layouts forobject-oriented programs Martin HirzelIBM Research SIGMETRICS 6/16/2007 Problem Object-oriented programs put datain objects. Caches and TLBs put data in blocks. Scattering objects over blocks causes cache/TLB misses. Misses cost time. Solution Most object-oriented languages use garbage collection. Garbage collection can move objects. To avoid misses, move objects to the right cache/TLB blocks. Simple, right? Cheney Copying GC BF: Breadth-first layout DF: Depth-first layout HI: Hierarchical layout AO: Allocation order layout SZ: Size segregation layout TH: Thread local layout Problem Which layout is best, and which is worst? How much does it matter in practice? How similar are the layouts? How much does it matter in the limit? Solutions? Appeal to intuition They can’t all be right! Formal Petrank/Rawitz showed hardness Simulation Who would believe those numbers? Brute-force Do you have a few person-years to spare? Avoiding Heisenberg Effects Object sorting garbage collection 32 Benchmarks % Mutator time overhead % Mutator miss rate increases Layout similarities and differences Estimated limit mutator time Benchmarks:largest avg. overhead Baseline:best observed Linear regression: Related work Conclusions Layouts matter little on average, but: Beware of the worst cases! Layout importance increases with SMP All layouts are sometimes best,sometimes worst AO has best average DF has most best-cases TH has best worst-cases Intel=SMT * o2 o1 o3 o4 o11 o9 o10 o5 o8 o6 o7 cache line TLB page o2 o1 o3 o4 o11 o9 o10 o5 o8 o6 o7 o11 o9 o10 o2 o1 o3 o4 o5 o8 o6 o7 scan free To-space Copied not yet scanned Copied scanned free scan scan free free scan From-space scan=free o9 o10 o2 o1 o3 o4 o5 o8 o6 o7 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Cheney 1970 Who? Siblings Why? Queue-based traversal How? Fenichel/Yochelson 1969 Who? Child-parent Why? Stack-based traversal How? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Moon 1984; Wilson et