Posted on 2008-03-15 14:46
ZelluX 閱讀(1031)
評(píng)論(6) 編輯 收藏 所屬分類:
Laboratory
包括各種paper, survey以及workshop上的講座等內(nèi)容
待讀：Memory Resource Management in VMware ESX Server,
Carl A. Waldspurger, OSDI 02
A Performance Study of General Purpose Applications on Graphics Processors, First Workshop on General Purpose Processing on GPU
The Google File System,
Sanjay Ghemawat, Howard Gobioff, and Shun-Tak Leung, SOSP 2003
Parallelization:1.
Automatic Parallelization for Graphic Processing Units,
Alan Leung, 6th Workshop on Compiler-Driven Performance
自動(dòng)識(shí)別Java程序中的并行化代碼，放到GPU上做，用到了RapidMind庫。
2.
Liquid SIMD: Abstracting SIMD Hardware using Lightweight Dynamic Mapping,
Nathan Clark.etc, University of Michigan & Cambridge, HPCA 07
粗讀
3.
Parallel Computing: What has changed lately?,
David B.Kirk, NVIDIA Corporation 2007
CUDA比CTM好用多了。。
4.
Data-Parallel Programming on the Cell BE and the GPU using the RapidMind Development Platform, GSPx Multicore Applications Conference
粗讀
5.
Mars: A MapReduce Framework on Graphics Processors香港科大、微軟、雅虎一起做的一個(gè)MapReduce庫，利用GPU進(jìn)行map操作會(huì)提升不少性能，不過reduce操作應(yīng)該會(huì)慢不少吧。
6.
A Survey of General-Purpose Computation on Graphics Hardware, EUROGRAPHICS 2005
寫得很全面，GPU計(jì)算的大致流程，相關(guān)算法，以及調(diào)試工具等方面都涉及了，里面提到的一些GPU缺陷現(xiàn)在也已經(jīng)被完善了。
7.
HMPP?: A Hybrid Multi-core Parallel Programming Environment,
CAPS entreprise
粗讀。和OpenMP蠻像的，有一些HMPP directives。
8.
Streamware: Programming General-Purpose Multicore Processors Using Streams,
Jayanth Gummaraju.etc, Stanford University, ASPLOS 08
精讀。適用于多種并行架構(gòu)（多核CPU、GPGPU）的編程平臺(tái)，對cache hierarchy的管理很出色（甚至通過這個(gè)平臺(tái)，一些程序跑在單處理器的條件下也能加速）。
Compilers:1.
Open Research Compiler (ORC): Beyond Version 1.0, PACT 02
主要看了其中的Loop Nested Optimization和Interprocedural Optimization
Virtulization:1.
SubVirt: Implementing malware with virtual machines,
Samuel T.King, etc., IEEE SP 06
閱讀筆記：
http://m.tkk7.com/zellux/archive/2008/05/05/198564.htmlhttp://m.tkk7.com/zellux/archive/2008/05/06/198693.html2.
Comptibility is Not Transparency: VMM Detection Myths and Realities,
Stanford University, VMWare, UBC/XenSource, Carnegie Mellon, Hot OS 07
聽過：1.
Accelerating Two-Dimensional PageWalks for Virtualized Systems,
Ravi Bhargava.etc小組例會(huì)上CC講的，Shadow Page Table開始就聽大不懂了 =_=。不過還是先放上來，占個(gè)位，恩。
2.
Inter-domain Socket Communications Supporting High Performance and Full Binary Compatibility on Xen,
Kangho Kim.etc, VEE 08
通過一套新的XWAY機(jī)制來提高Xen中不同Domain互相之間網(wǎng)絡(luò)傳輸速率。不過僅限于TCP協(xié)議。
3.
Dispersing Proprietary Applications as Benchmarks through Code Mutation,
Luk Van Ertvelde & Lieven Eeckhout, ASPLOS 08
很好玩的一篇paper。現(xiàn)在的benchmark跑的程序通常都是開源的或是免費(fèi)的軟件，但是商業(yè)軟件就會(huì)涉及到版權(quán)問題。
這篇paper講的是如何通過二進(jìn)制代碼的突變(mutation)達(dá)到更改軟件功能，但又不影響軟件跑在benchmark上時(shí)所耗的時(shí)間。
4.
Threads Cannot Be Implemented As a Library, ACM SIGPLAN 2005
?舉例說明了用庫實(shí)現(xiàn)線程（如pthread）的一些缺陷