Download presentation
Presentation is loading. Please wait.
Published byNeil Benson Modified over 9 years ago
1
PVM and MPI What Else is Needed For Cluster Computing? Al Geist Oak Ridge National Laboratory www.csm.ornl.gov/~geist DAPSYS/EuroPVM-MPI Balatonfured, Hungary September 11, 2000
2
EuroPVM-MPI EuroPVM-MPI Dedicated to the hottest developments of PVM and MPI PVM and MPI are the most used tools for parallel programming The hottest trend driving PVM and MPI today is PC clusters running Linux and/or Windows This talk will look at gaps in what PVM and MPI provide for Cluster Computing. What role the GRID may play and What is happening to fill the gaps…
3
New release this summer – PVM 3.4.3 includes: Optimized msgbox routines –More scalable, more robust New Beowulf-linux port –Allows clusters to be behind firewalls yet work together Smart virtual machine startup –Automatically determines the reason for “Can’t start pvmd” Works with Windows2000 –Installshield version available –Improved Win32 communication performance PVM Latest News: New Third party PVM Software: PythonPVM 0.9, JavaPVM, interface PVM port using SCI interface
4
198994909697992000 PVM-1 MPI-1MPI-2I-MPI PVM-2PVM-3PVM-3.4 Harness Wide-area GRID experiments Ten Years of Cluster Computing Building a Cluster Computing Environment for 21 st Century Networks of Workstations PC Clusters
5
0.1 1 10 100 1000 10000 100000 1000000 Jun-93 Nov-94 Jun-96 Nov-97 Jun-99 Nov-00 Jun-02 Nov-03 Jun-05 Nov-06 Jun-08 Nov-09 Performance [GFlop/s] N=1 N=500 N=10 1 TFlop/s 1 PFlop/s 2005 Entry at 1TFlop/s 2010 Peak at 1PFlop/s ASCI IBM Compaq TOP500 Trends – Next 10 years Even the largest machines are clusters http://www.netlib.org/benchmark/top500.html
6
PC Clusters are cost effective from a hardware perspective Many Universities and companies can afford 16 to 100 nodes. System administration is an overlooked cost: people to maintain cluster software written for each cluster higher failure rates for COTS Presently there is lack of tools for managing large clusters www.csm.ornl.gov/torc Trend in Affordable PC clusters
7
C3 Command line cluster toolset Cluster Computing Tools cl_pushimage() – push system image across cluster cl_shutdown() – shutdown specified nodes cl_push() – push files/directories across cluster cl_rm() – remove files from multiple nodes cl_get() – gather cluster files to one location cl_ps() – returns results of multi-node ps cl_kill() – kill application across entire cluster cl_exec() – execution of any command across specified nodes Functions C3 is a command line based toolset for system administration and user level operations on a single cluster. C3 functions may also be called in a program. C3 is multithreaded and each function executes in parallel. only executable by sysadmin Software available www.csm.ornl.gov/torc
8
VTK ported to Linux Clusters –Visualization toolkit making use of C3 cluster package AVS Express ported to PC Cluster –Expensive but standard package asked for by apps. –Requires AVS Site license to eliminate the cost of individual node licenese Cumulvs plug-in for AVS Express –Combines the interactive visualization and computational steering of Cumulvs with the visualization tools of AVS. Visualization using Clusters Lowering the Cost of High Performance Graphics
9
M3C tool suite Cluster Computing Tools Reserve nodes within or across clusters Submit Job to queue system Monitor nodes of cluster – also adding a PAPI interface Install software on selected nodes Reboot, shutdown, add user, etc. to selected cluster nodes. Display properties of nodes Growing list of plug-in modules Suite of user interface tools for system administration and simultaneous monitoring of multiple PC clusters. Written in Java – web based remote access. www.csm.ornl.gov/torc
10
OSCAR Bring uniformity to cluster creation and use Make clusters more broadly acceptable Foster commercial versions of the cluster software Goals OSCAR is a collection of the best known software for building, programming, and using clusters. The collection effort is lead by a national consortium which includes: IBM, SGI, Intel, ORNL, NCSA, MCS Software. Other vendors invited. www.csm.ornl.gov/oscar National Consortium for Cluster software For more details see Stephen Scott (talk Monday 12:00 DAPSYS track)
11
ORNL M C Tool Architecture designed to work both within and across organizations 3 URL CGI C3 Scripts/monitors URL CGI Custom Scripts URL CGI Third-party Scripts GUI proxy URL front-end back-end M C 3 Java applet Based GUI cluster 1cluster 2 cluster 1 Interface thru XML files M3C proxy allows one sysadmin to monitor and update multiple clusters M3C GUI allows user to submit and monitor jobs ORNL UTK SDSC
12
Information services Uniform naming, locating, and allocating distirbuted resources Data management and access Single log-on security GRID “Ubiquitous” Computing GRID Forum is helping define higher level services MPI and PVM are often seen as lower level capabilities that GRID frameworks support. GlobusNetSolveCondor LegionNeosSinRG
13
Cumulvs – collaborative computational steering Recent Highlights Release of new version Development of CAVE viewer Works w/ PNL global arrays Made CCA compliant Cumulvs was the initial reason we started Harness. Collective port A DOE effort to provide a standard for interoperability of high performance components developed by many different groups in different languages or frameworks. Common Component Architecture http://z.ca.sandia.gov/~cca-forum/port-spec
14
HARNESS Exploring New Capabilities in Heterogeneous Distributed Computing www.epm.ornl.gov/harness Parallel Plug-in environment Extend the concept of a plug-in to the parallel computing world. Dynamic with no restrictions on functions. Distributed peer-to-peer control No single point of failure unlike typical client/server models. Multiple distributed virtual machines merge/split Provide a means for short-term sharing of resources and collaboration between teams. Building on our experience and success with PVM create a fundamentally new heterogeneous virtual machine based on three research concepts: Goal
15
Host D Host C Host B Host A Virtual Machine Operation within VM uses Distributed Control process control user features HARNESS daemon Customization and extension by dynamically adding plug-ins Component based daemon Merge/split with other VMs Another VM HARNESS Virtual Machine Scalable Distributed control and Component based Daemon
16
HARNESS Latest News Provide a practical environment and illustrate extensibility Harness Core (beta release ready)(see talk Monday 16:30 Track 1) Task library and Harness Daemon software Provides API to load, unload plug-ins and distributed control. PVM Plug-in (stalled for summer now back on track) Provides PVM API veneer to support exiting PVM applications. Fault tolerant MPI plug-in (see talk Monday 16:50 Track 1) Provides MPI API for 30 most used functions. Semantics adjusted to allow recovery from corrupted communicator. VIA communication plug-in (looking at multi-interface transfer) To illustrate how different low level communication plug-ins can be used within Harness. And to provide high performance
17
Parallel Plug-in Research For Heterogeneous Distributed Virtual Machine Tough Research problems include: Heterogeneity (has delayed ‘C’ H-core development.) Synchronization - Dynamic installation Interoperation between same plug-in on different tasks between task plug-in and daemon plug-in between daemon plug-ins Partial success One research goal is to understand and implement a dynamic parallel plug-in environment. provides a method for many users to extend Harness in much the same way that third party serial plug-ins extend Netscape, Photoshop, and Linux.
18
Fault Tolerant MPI Motivation As application and machine sizes grow the MTBF is less than the application run time. MPI standard is based on a static model so any decrease in tasks leads to corrupted communicator (MPI_COMM_WORLD). Develop MPI plugin that takes advantage of Harness robustness to allow a range of recovery alternatives to an MPI application. Not just another MPI implementation. FT-MPI follows the syntax of MPI standard Communication performance on par with MPICH Presently uses PVM3.4.3 fault recovery until Harness is ready)
19
Fault Tolerant MPI Recovery requires MPI semantic changes Key step to MPI recovery is creating a communicator that app can use to continue. Accomplished by modifying the semantics of two MPI functions. MPI_COMM_CREATE ( comm, group, newcomm) MPI_COMM_SPLIT ( comm, color, key, newcomm) Creates a new communicator that contains all surviving processes Allows MPI_COMM_WORLD to be specified as both input and output communicator.
20
No single point (or set of points) of failure for Harness. It survives as long as one member still lives. All members know the state of the virtual machine, and their knowledge is kept consistent w.r.t. the order of changes of state. (Important parallel programming requirement!) No member is more important than any other (at any instant) i.e. here isn’t a pass-around “control token” Symmetric Peer-to-Peer Distributed Control Characteristics
21
Harness kernels on each host have arbitrary priority assigned to them (new kernels are always given the lowest priority) Virtual machine A task on this host requests a new host be added VM state held by each kernel 2. Each adds request to a list of pending changes 1. Send host/T#/data to neighbor in ring 3. Distributed Control Harness Two Phase Arbitration
22
Supports multiple simultaneous updates Harness Distributed Control Control is Scalable, Asynchronous, and Parallel add host Fast host delete or recovery from fault Parallel recovery from multiple host failures Supports fast host adding Scalable Design 1<=S<=P
23
Follow the links from my Web site www.csm.ornl.gov/~geist For more Information Also - Copy of these slides
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.