DySectAPI: Scalable Prescriptive Debugging

Nicklas Bo Jensen; Sven  Karlsson; Niklas  Quarfot Nielsen; Gregory L. Lee; Dong H. Ahn; Matthew  Legendre; Martin Schulz

DySectAPI: Scalable Prescriptive Debugging

Nicklas Bo Jensen, Sven Karlsson, Niklas Quarfot Nielsen, Gregory L. Lee, Dong H. Ahn, Matthew Legendre, Martin Schulz

Research output: Contribution to conference › Conference abstract for conference › Research › peer-review

274 Downloads (Pure)

Abstract

We present the DySectAPI, a tool that allow users to construct probe trees for automatic, event-driven debugging at scale. The traditional, interactive debugging model, whereby users manually step through and inspect their application, does not scale well even for current supercomputers. While lightweight debugging models scale well, they can currently only debug a subset of bug classes. DySectAPI fills the gap between these two approaches with a novel user-guided approach. Using both experimental results and analytical modeling we show how DySectAPI scales and can run with a low overhead on current systems.

Original language	English
Publication date	2014
Number of pages	2
Publication status	Published - 2014
Event	International Conference for High Performance Computing, Networking, Storage and Analysis, SC14 - New Orleans, United States Duration: 16 Nov 2014 → 21 Nov 2014 http://sc14.supercomputing.org/

Conference

Conference	International Conference for High Performance Computing, Networking, Storage and Analysis, SC14
Country/Territory	United States
City	New Orleans
Period	16/11/2014 → 21/11/2014
Internet address	http://sc14.supercomputing.org/

Access to Document

Extended abstractFinal published version, 153 KB

http://sc14.supercomputing.org/sites/all/themes/sc14/files/archive/tech_poster/tech_poster_pages/post237.html

OpenUrl availability

Full text

Cite this

Jensen, N. B., Karlsson, S., Quarfot Nielsen, N., Lee, G. L., Ahn, D. H., Legendre, M., & Schulz, M. (2014). DySectAPI: Scalable Prescriptive Debugging. Abstract from International Conference for High Performance Computing, Networking, Storage and Analysis, SC14, New Orleans, United States. http://sc14.supercomputing.org/sites/all/themes/sc14/files/archive/tech_poster/tech_poster_pages/post237.html

@conference{a0745a7578c44c5d94adff0204c7d53e,

title = "DySectAPI: Scalable Prescriptive Debugging",

abstract = "We present the DySectAPI, a tool that allow users to construct probe trees for automatic, event-driven debugging at scale. The traditional, interactive debugging model, whereby users manually step through and inspect their application, does not scale well even for current supercomputers. While lightweight debugging models scale well, they can currently only debug a subset of bug classes. DySectAPI fills the gap between these two approaches with a novel user-guided approach. Using both experimental results and analytical modeling we show how DySectAPI scales and can run with a low overhead on current systems.",

author = "Jensen, {Nicklas Bo} and Sven Karlsson and {Quarfot Nielsen}, Niklas and Lee, {Gregory L.} and Ahn, {Dong H.} and Matthew Legendre and Martin Schulz",

year = "2014",

language = "English",

note = "International Conference for High Performance Computing, Networking, Storage and Analysis, SC14, SC ; Conference date: 16-11-2014 Through 21-11-2014",

url = "http://sc14.supercomputing.org/",

}

Jensen, NB, Karlsson, S, Quarfot Nielsen, N, Lee, GL, Ahn, DH, Legendre, M & Schulz, M 2014, 'DySectAPI: Scalable Prescriptive Debugging', International Conference for High Performance Computing, Networking, Storage and Analysis, SC14, New Orleans, United States, 16/11/2014 - 21/11/2014. <http://sc14.supercomputing.org/sites/all/themes/sc14/files/archive/tech_poster/tech_poster_pages/post237.html>

TY - ABST

T1 - DySectAPI: Scalable Prescriptive Debugging

AU - Jensen, Nicklas Bo

AU - Karlsson, Sven

AU - Quarfot Nielsen, Niklas

AU - Lee, Gregory L.

AU - Ahn, Dong H.

AU - Legendre, Matthew

AU - Schulz, Martin

PY - 2014

Y1 - 2014

N2 - We present the DySectAPI, a tool that allow users to construct probe trees for automatic, event-driven debugging at scale. The traditional, interactive debugging model, whereby users manually step through and inspect their application, does not scale well even for current supercomputers. While lightweight debugging models scale well, they can currently only debug a subset of bug classes. DySectAPI fills the gap between these two approaches with a novel user-guided approach. Using both experimental results and analytical modeling we show how DySectAPI scales and can run with a low overhead on current systems.

AB - We present the DySectAPI, a tool that allow users to construct probe trees for automatic, event-driven debugging at scale. The traditional, interactive debugging model, whereby users manually step through and inspect their application, does not scale well even for current supercomputers. While lightweight debugging models scale well, they can currently only debug a subset of bug classes. DySectAPI fills the gap between these two approaches with a novel user-guided approach. Using both experimental results and analytical modeling we show how DySectAPI scales and can run with a low overhead on current systems.

M3 - Conference abstract for conference

T2 - International Conference for High Performance Computing, Networking, Storage and Analysis, SC14

Y2 - 16 November 2014 through 21 November 2014

ER -

DySectAPI: Scalable Prescriptive Debugging

Abstract

Conference

Access to Document

OpenUrl availability

Fingerprint

Cite this