Sökning: id:"swepub:oai:DiVA.org:kth-186850" >
Backlog bound analy...
Backlog bound analysis for virtual-channel routers
-
- Zhao, Xueqian (författare)
- KTH,Elektronik och Inbyggda System
-
- Lu, Zhonghai (författare)
- KTH,Elektronik och Inbyggda System
-
(creator_code:org_t)
- Institute of Electrical and Electronics Engineers (IEEE), 2015
- 2015
- Engelska.
-
Ingår i: 2015 IEEE Computer Society Annual Symposium on VLSI. - : Institute of Electrical and Electronics Engineers (IEEE). - 9781479987191 ; , s. 422-427
- Relaterad länk:
-
https://urn.kb.se/re...
-
visa fler...
-
https://doi.org/10.1...
-
visa färre...
Abstract
Ämnesord
Stäng
- Backlog bound analysis is crucial for predicting buffer sizing boundary in on-chip virtual-channel routers. However, the complicated resource contention among traffic flows makes the analysis difficult. Because conventional simulation-based approaches are generally incapable of investigating the worst-case scenarios for the backlog bounds, we propose a formal analysis technique. We identify basic buffer use scenarios and propose corresponding analysis models to formally deduce per-buffer backlog bound using network calculus. A topology independent analysis technique is developed to convey the per-buffer backlog bound analysis step by step. We further develop an algorithm to automate the analysis procedure with polynomial complexity. A case study shows how to apply the technique and the analytical bounds are tight.
Ämnesord
- TEKNIK OCH TEKNOLOGIER -- Elektroteknik och elektronik (hsv//swe)
- ENGINEERING AND TECHNOLOGY -- Electrical Engineering, Electronic Engineering, Information Engineering (hsv//eng)
Nyckelord
- Network calculus
- Network on chip
- Performance analysis
- Calculations
- Complex networks
- Network-on-chip
- VLSI circuits
- Analytical bounds
- Polynomial complexity
- Resource contention
- Simulation based approaches
- Topology-independent
- Worst case scenario
- Routers
Publikations- och innehållstyp
- ref (ämneskategori)
- kon (ämneskategori)
Hitta via bibliotek
Till lärosätets databas