ACM Home Page
Please provide us with feedback. Feedback
Interactive presentation: System-level process variation driven throughput analysis for single and multiple voltage-frequency island designs
Full text PdfPdf (181 KB)
Source Design, Automation, and Test in Europe archive
Proceedings of the conference on Design, automation and test in Europe table of contents
Nice, France
SESSION: Synthesis at system and architectural levels table of contents
Pages: 403 - 408  
Year of Publication: 2007
ISBN:978-3-9810801-2-4
Authors
Siddharth Garg  Carnegie Mellon University
Diana Marculescu  Carnegie Mellon University
Sponsors
: IEEE Council on Electronic Design Automation (CEDA)
SIGDA: ACM Special Interest Group on Design Automation
: The EDA Consortium
EDAA : European Design and Automation Association
RAS : RAS
: The IEEE Computer Society TTTC
: ECSI
Publisher
EDA Consortium  San Jose, CA, USA
Bibliometrics
Downloads (6 Weeks): 1,   Downloads (12 Months): 6,   Citation Count: 3
Additional Information:

abstract   references   cited by   collaborative colleagues  

Tools and Actions: Review this Article  

ABSTRACT

Manufacturing process variations are the primary cause of timing yield loss in aggressively scaled technologies. In this paper, we analyze the impact of process variations on the throughput (rate) characteristics of embedded systems comprised of multiple voltage-frequency islands (VFIs) represented as component graphs. We provide an efficient, yet accurate method to compute the throughput of an application in a probabilistic scenario and show that systems implemented with multiple VFIs are more likely to meet throughput constraints than their fully synchronous counterparts. The proposed framework allows designers to investigate the impact of architectural decisions such as the granularity of VFI partitioning on their designs, while determining the likelihood of a system meeting specified throughput constraints. An implementation of the proposed framework is accurate within 1.2% of Monte Carlo simulation while yielding speed-ups ranging from 78X-260X, for a set of synthetic benchmarks. Results on a real benchmark (MPEG-2 encoder) show that a nine clock domain implementation gives 100% yield for a throughput constraint for which a fully synchronous design only yields 25%. For the same throughput constraint, a three clock domain architecture yields 78%.


REFERENCES

Note: OCR errors may be found in this Reference List extracted from the full text article. ACM has opted to expose the complete List rather than only correct and linked references.

 
1
 
2
3
4
 
5
6
7
8
 
9
 
10
R. M. Karp. A characterization of the minimum cycle mean in a digraph. Discrete Math, 23, 1978.
11
 
12

Collaborative Colleagues:
Siddharth Garg: colleagues
Diana Marculescu: colleagues