DidaWiki

23/02/2010 Introduction to the course — Aim and purpose, program, structure, rules and kind of final tests.

24/02/2010 Grid computing, (I) — Large-scale problems in research and prodution environments. How to approach these problems. Preliminary definitions: resources, protocols, services, APIs and SDKs. A simple example: web services and their protocols.

26/02/2010 Grid Computing, (II) — Virtual Organizations. The Grid vision and its requirements. The Grid architecture: fabric, connectivity, resource, collective and application layers. Using the Grid: scenarios and examples. Open Grid Service Architecture and its capabilities. The eight fallacies of Grid computing.

02/03/2010 Message Passing Interface, Part 1 — The MPI standard, brief history and purpose. Message Passing programming and portability, functional and performance portability. The library approach of MPI. MPI basic concepts: communicators, point to point primitives, collective primitives, datatypes; communication semantics in MPI: envelopes, local/global completion, (non)blocking operations, communication modes. Opaque Objects. Primitive datatypes.

03/03/2010 MPI, Part 2

05/03/2010 MPI, Part 3

09/03/2010 Grid Computing (III) — The Globus Project; Public Key Infrastructure (concepts). Grid Security Infrastructure. Certificates. Single Sign On and Delegation.

10/03/2010 Grid Computing (IV) — Grid Information Services. Lightweight Directory Access Protocol (concepts). Monitoring and Discovery Service. IP, GRIS and GIIS. Grid Information Models: MDS-2 and GLUE schemata

12/03/2010 MPI, Part 4

16/03/2010 Virtualization technology (I) — Basics of virtualization at the program language, operating system and machine level. Quick survey of Pascal P-code, Java bytecode and .Net CLI. Hardware virtualization by emulation. Read papers by J.E.Smith listed in the course bibliography

17/03/2010 Virtualization technology (II), MPI lab — Quick tour of full-system virtualization: hosted and bare-metal hypervisors; protection rings in the x86 architecture; virtualization and paravirtualization, hardware-assisted virtualization; containers and hosted virtualization. Elementary programs in MPI.

19/03/2010 Grid Computing (V) — HPC Resource Management; Grid Resource Management; Gatekeeper and Job Manager; Data Management; GASS and GridFTP; Replica Catalog and Replica Management Services

23/03/2010 Map & reduce: the programming model — Problem characterization; Map Fold in LISP; Programming model: mappers and reducers; Programming model: partitioners and combiners; Example and data flow

24/03/2010 Distributed File Systems: GFS and HDFS — Problem characterization; Blocks, Name nodes and Data Nodes; Master/server architecture; Master Server (namenode), Chunk Servers (datanode) protocols and responsabilities; Anatomy of a read; Anatomy of a write; Benchmarks

26/03/2010 Hadoop Lab — Hadoop installation and setup; Single mode and pseudodistributed mode configuration; Grep application

Electoral break, Easter break

09/04/2010 MPI Lab

13/04/2010 Hadoop Lab — Word Count application, versions with both old and new Hadoop APIs; API usage; using a large number of files

14/04/2010

16/04/2010 MPI Lab — Simple examples: buffer ping pong, square and triangular matrices with transpose on send/receive.

20/04/2010 Application Examples Mandelbrot; brief survey of KDD and Data Mining; the K-means algorithm

21/04/2010 K-means + MPI Lab

23/04/2010 MPI Lab

27/04/2010 XtreemOS part I

28/04/2010

30/04/2010

04/05/2010 (see the CPA course page)

05/05/2010 (see the CPA course page)

07/05/2010 (see the CPA course page)

11/05/2010 (see the CPA course page)

12/05/2010 (see the CPA course page)

14/05/2010 The ASSIST environment, part I — relationship with other skeleton-based parallel programming tools; basic concepts, flexible-skeletons, multi-language support and heterogeneous platforms; sequential code encapsulation, stream-oriented communication, generic graphs, the parmod as an abstract “wildcard” skeleton

18/05/2010 The ASSIST environment, part II — parmod semantics and usage rationale; structure and semantic options available to the programmer: input and output section, virtual processes, distribution, collection; mapping of VPs over concrete architectures.

19/05/2010 The ASSIST environment, part III — support of heterogeneity and dynamic adaptivity at run-time; structure and performance contract directed deployment.

21/05/2010 Presentation of project proposals