TotalView Debugger On Blue Gene

67 %
33 %
Information about TotalView Debugger On Blue Gene
Technology

Published on June 16, 2009

Author: Totalviewtech

Source: slideshare.net

Description

Totalview Debugger on Blue Gene

Scalable Debugging with TotalView on Blue Gene  John DelSignore, CTO TotalView Technologies

Agenda • TotalView on Blue Gene – A little history – Current status • Recent TotalView improvements – ReplayEngine (reverse debugging) – Remote Display – TotalView Script (batch debugging) • Future work – BG/* – Heterogeneous systems – Many core, transactional memory, speculative execution – Peta­scale debugging TotalView Technologies – Confidential and Proprietary – Plans Subject to Change without Notice www.totalviewtech.com 2

Supported Blue Gene  Architectures and Compilers • Blue Gene/L and Blue Gene/P • Languages / Compilers – C/C++, Fortran, Assembly – GNU Compilers – IBM Compilers – IBM OpenMP (on BG/P) • Parallel Environments – IBM MPI  – IBM OpenMP (on BG/P) – Pthreads (BG/P) • Runtime linking/loading (BG/P) – Shared libraries – Dynamically loaded shared libraries TotalView Technologies – Confidential and Proprietary – Plans Subject to Change without Notice www.totalviewtech.com 3

Blue Gene Architecture • TotalView client (GUI/CLI)  runs on the Front End node • Client communicates with  the TotalView debugger  servers running on the I/O  nodes via a socket • The debugger servers  communicate with the  CIOD to control processes  and threads running on the  Compute nodes • Fan­out ratios (CNs/server) – BG/L: 32­64, 2 cores/CN,  128 threads/server – BG/P:128­256, 4 cores/CN,  1024 threads/server – Ratio increasing (8K thr/svr?) – Parallelize server operation TotalView Technologies – Confidential and Proprietary – Plans Subject to Change without Notice www.totalviewtech.com 4

TotalView Blue Gene/L Support • TotalView involvement since 2003 • Support for Blue Gene/L since 2005 • Debugging interfaces developed via close  collaboration with IBM • Used on DOE/NNSA/LLNL's  Blue Gene/L system  containing 212 K cores – Heap memory debugging support added – Blue Gene/L scaling and performance tuning project – TotalView has debugged jobs as large as 8,192 processes  (LLNL) • Work on Blue Gene/L facilitated Blue Gene/P  support TotalView Technologies – Confidential and Proprietary – Plans Subject to Change without Notice www.totalviewtech.com 5

TotalView Blue Gene/P Support • Blue Gene/P supported since Q4 2007 • Continued close collaboration with IBM to  develop multi­threaded debugging interfaces • Support for shared libraries and dynamically  loaded libraries • Scalability improvements • TotalView has debugged jobs as large as 32K  (Jülich) TotalView Technologies – Confidential and Proprietary – Plans Subject to Change without Notice www.totalviewtech.com 6

TotalView Blue Gene/P Sites • Currently running at over  30 sites in Germany,  France, UK, and US, including – Argonne – Boston University – Daresbury – IDRIS – Jülich – LLNL – Max Planck – ORNL – Princeton University – Rensselaer Polytechnic Institute • Jülich workshop, March 08 • Argonne workshop, May 08 TotalView Technologies – Confidential and Proprietary – Plans Subject to Change without Notice www.totalviewtech.com 7

Recent TotalView Improvements on Blue Gene and Linux • Remote Display – Run a remote version of the TotalView GUI… – …display it locally, with fast, interactive performance – Easy, fast, secure • tvscript – Simplifies debugging batch jobs – Event/action paradigm – Configurable • ReplayEngine – Step execution back in time – Uses reverse debugging technology – Linux x86 and x86­64 (currently only) TotalView Technologies – Confidential and Proprietary – Plans Subject to Change without Notice www.totalviewtech.com 8

Remote Display • Presents a window on your  machine that will display  TotalView executing on a  remote system • Two components:  – Client, runs on the local  system, available for   Linux x86, x86­64  Windows XP, Vista – Server, which runs on any  system supported by  TotalView, invisibly  managing the connections  between the host and client • The Client also provides for  submission of jobs to  batch queuing systems  PBS Pro and LoadLeveler TotalView Technologies – Confidential and Proprietary – Plans Subject to Change without Notice www.totalviewtech.com 9

Batch Scripting • Designed for debugging in a batch environment • tvscript lets you define the events to act on, the actions to  take when an event occurs • Typical events – Action point (e.g., breakpoint) – Memory error (e.g., malloc returns 0, guard block corruption) – Errors (e.g., SEGV, FPE) • Typical actions – Display a backtrace – List memory leaks – Print variables and arrays • Configurable – Supports external script files – Allows generation of even more complex actions and events TotalView Technologies – Confidential and Proprietary – Plans Subject to Change without Notice www.totalviewtech.com 10

Replay Engine • Intuitive user interface, integrated with TotalView Step forward over functions Step backward over functions Step forward into functions Step backward into functions Advance forward out of current  Advance backward out of  current Function, after the call Function, to before the call Advance forward to selected line  Advance backward to selected line Advance forward to “live” session TotalView Technologies – Confidential and Proprietary – Plans Subject to Change without Notice www.totalviewtech.com 11

Possible Future Blue Gene Work • BG/* support – Support future generations of Blue Gene • Fast conditional breakpoints/watchpoints – Expressions compiled/patched into target, excute in parallel,  about 10usecs/expression • Asynchronous thread control – Thread barrier breakpoint, thread single stepping • User programmable visual data – Allows user define complex data access function • Debugging optimized code • Post­mortem debugging • Fast DLL debugging interface • LLNL collaboration for scalable subset attach – Integrates with lightweight tools such as STAT TotalView Technologies – Confidential and Proprietary – Plans Subject to Change without Notice www.totalviewtech.com 12

Possible Other Future Work • Scalability/performance – Continue scalability and performance improvements – Tree­based infrastructure for logarithmic scaling – Peta­scale debugging – Hundreds of thousands of threads • Heterogeneous systems – IBM Roadrunner (x86­64/Cell) – GPUs • Emerging technologies – Many core – Transactional memory – Speculative execution TotalView Technologies – Confidential and Proprietary – Plans Subject to Change without Notice www.totalviewtech.com <number>

Questions?   More Information • Blue Gene Technical Development Interest Group Contact chris.gottbrath@totalviewtech.com – • Technical support  support@totalviewtech.com – • BG LLNL case study www.totalviewtech.com/pdf/case_study_scientific_computing.pdf  – • Customer training or webinars contacttraininggroup@totalviewtech.com – • Web site  – www.totalviewtech.com TotalView Technologies – Confidential and Proprietary – Plans Subject to Change without Notice www.totalviewtech.com <number>

Add a comment

Related presentations

Related pages

TotalView Debugger on Blue Gene P | Argonne Leadership ...

peter.thompson@totalviewtech.com. Institution: TotalView Technologies
Read more

Debugging with TotalView on the Blue Gene Q - alcf.anl.gov

Ed Hinkel, Sr Sales Engineer Rogue Wave Software Debugging with TotalView on the Blue Gene Q
Read more

Blue Gene/L Workshop TotalView on Blue Gene/L Reno, NV ...

TotalView on Blue Gene/L John DelSignore Chief Architect Blue Gene/L Workshop Reno, NV October, 2003. Outline of Today’s Talk ... TotalView Debugger ...
Read more

Rogue Wave Releases TotalView with Blue Gene/Q Support

Rogue Wave Releases TotalView with Blue Gene/Q ... version of the TotalView debugger for porting codes ... choice for debugging on the Blue Gene/Q ...
Read more

TotalView Debugger Stands Above with Support for Blue Gene/Q

In this video from SC12, Scott Lasica from Rogue Wave Software discusses the company's new Totalview programming tools for IBM Blue Gene/Q ...
Read more

CoCoBoard Mail - Boston University

The totalview debugger is now available for use on the Blue Gene under a demonstration license which will expire at the end of May. On Thursday, May 3 ...
Read more

Content Standard Tools IBM System Blue Gene/P Specifics ...

– TotalView Debugger – Allinea DDT. ... IBM System Blue Gene-P - Debugging - 1-2-0 Author: tallet Created Date: 6/23/2010 2:14:29 PM
Read more

TotalView: Graphical Debugger, Memory Leak Detection and ...

... (including the Cray and Blue Gene platforms), Linux PowerLE, UNIX, Mac OS X, ... “We rely on TotalView for all of our parallel debugging needs.
Read more

SUR Blue Gene - ScorecWiki - scorec.rpi.edu

This document is a repository for basic usage information pertaining to RPI's SUR Blue Gene ... We have licensed the Totalview debugger for the Blue Gene ...
Read more