Difference between revisions of "Cluster:LLK Poster Structure"
Jump to navigation
Jump to search
| Line 1: | Line 1: | ||
==Abstract== | ==Abstract== | ||
Network latency continues to be a bottleneck for many computational science applications on Beowulf clusters. Using open source tools, we develop technology for making nanosecond precision measurements between timing points within the 2.6 Linux kernel. Using this technology, in conjunction with low-level and application benchmarks, we establish that there is significant network latency in the kernel and identify its origin. Lastly, we survey software-based latency reduction technologies and make recommendations based on our measurements. ([[Cluster:llk-siam-pp06-abstract|original]]) | Network latency continues to be a bottleneck for many computational science applications on Beowulf clusters. Using open source tools, we develop technology for making nanosecond precision measurements between timing points within the 2.6 Linux kernel. Using this technology, in conjunction with low-level and application benchmarks, we establish that there is significant network latency in the kernel and identify its origin. Lastly, we survey software-based latency reduction technologies and make recommendations based on our measurements. ([[Cluster:llk-siam-pp06-abstract|original]]) | ||
| + | |||
| + | '''Keywords:''' | ||
| + | |||
==The Problem== | ==The Problem== | ||
===Overview=== | ===Overview=== | ||
| + | |||
===Motivation=== | ===Motivation=== | ||
| + | |||
===Survey/Current State of Affairs=== | ===Survey/Current State of Affairs=== | ||
Places to look: | Places to look: | ||
| Line 11: | Line 16: | ||
* Citeseer | * Citeseer | ||
* Google Scholar | * Google Scholar | ||
| + | |||
==Methodology== | ==Methodology== | ||
===Measurement (Techniques)=== | ===Measurement (Techniques)=== | ||
| − | + | * TCP/IP stack time point placement | |
| + | * tp_timer | ||
| + | * high res timing w/counters | ||
| + | |||
===Verification and Validation=== | ===Verification and Validation=== | ||
| − | ===Analysis=== | + | |
| + | ===(Analysis) Tools=== | ||
| + | * /proc -> database | ||
| + | * PHP/gnuplot visualization | ||
| + | |||
===Benchmarks=== | ===Benchmarks=== | ||
| + | |||
==Test Results== | ==Test Results== | ||
(tabular, incl. test environment) | (tabular, incl. test environment) | ||
| + | |||
==Summary/Discussion/Conclusions== | ==Summary/Discussion/Conclusions== | ||
| + | |||
==Future Work== | ==Future Work== | ||
| + | |||
==References== | ==References== | ||
Revision as of 11:32, 28 January 2006
Contents
Abstract
Network latency continues to be a bottleneck for many computational science applications on Beowulf clusters. Using open source tools, we develop technology for making nanosecond precision measurements between timing points within the 2.6 Linux kernel. Using this technology, in conjunction with low-level and application benchmarks, we establish that there is significant network latency in the kernel and identify its origin. Lastly, we survey software-based latency reduction technologies and make recommendations based on our measurements. (original)
Keywords:
The Problem
Overview
Motivation
Survey/Current State of Affairs
Places to look:
- Beowulf archives
- ACM
- IEEE
- Citeseer
- Google Scholar
Methodology
Measurement (Techniques)
- TCP/IP stack time point placement
- tp_timer
- high res timing w/counters
Verification and Validation
(Analysis) Tools
- /proc -> database
- PHP/gnuplot visualization
Benchmarks
Test Results
(tabular, incl. test environment)
