Retiming for Synchronous Data Flow Graphs

Liveris, N.; Lin, Cuiying; Wang, J.; Zhou, Hai; Banerjee, P.

doi:10.1109/aspdac.2007.358032

Cited by 13 publications

(15 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Liveris et al [15] were the first to explore the relation between an SDFG and its equivalent HSDFG to achieve high efficiency in retiming SDFGs. We further elaborate on this inter-relation.…”

Section: Basic Idea Of Our Methodsmentioning

confidence: 99%

“…Following [15], potentially optimal IPs are checked decreasingly until we find one that is not feasible.…”

Section: Basic Idea Of Our Methodsmentioning

confidence: 99%

“…Although a binary search has a lower theoretical complexity than a linear search, a binary search method for optimal retiming tends to test more than one potential optIP that is not feasible, while a linear search method, searching potential optIPs decreasingly, only tests one potential optIP that is not feasible. So, in line with the approach in [15], we use a linear search to find the optimal retiming, called sdfOPT, shown in Algorithm 4. run Algorithm sdfFEAS(G r , optIP − 1) to determine whether a feasible retiming exists 7: if a feasible retiming r ′ exists then 8: G r = r ′ (G r ); 9: r+ = r ′ ; 10: get IP of G r from sdfIP(G r ); 11: optIP = IP 12: else 13: return r 14: end if 15: end while 16: return r Since initIP is a feasible IP, sdfOPT begins with dip = initIP − 1, using sdfFEAS to check whether a feasible retiming exists for G with this dip. If so, it stores this retiming and computes the new IP of the retimed graph.…”

Section: Optimal Retimingmentioning

confidence: 99%

“…But the paper says nothing about necessity. Liveris et al [15] present an optimal retiming method that computes both the iteration period and retiming function on SDFGs but it does not deal with the feasible retiming problem, although of course it can be adapted straightforwardly. They use a complex termination condition and prove that it is sufficient for an optimal retiming to be found, but no evidence is given that it is sharp.…”

mentioning

confidence: 99%

“…The methods for feasible retiming in our comparison include the traditional method [13], 'byHSDF', O'Neil's algorithm [3], 'O'Neil01', the method in [14], 'Zhu10', and our new method, 'sdfFEAS'. The methods for optimal retiming include Liveris's algorithm [15] , 'Liveris07', and our new method, 'sdfOPT'. The experimental results, based on hundreds of synthetic SDFGs and several realistic SDFG models, show that our new methods lead to a significant improvement compared to earlier methods.…”

mentioning

confidence: 99%

See 4 more Smart Citations

Efficient Retiming of Multirate DSP Algorithms

Zhu

Basten

Geilen

et al. 2012

IEEE Trans. Comput.-Aided Des. Integr. Circuits Syst.

View full text Add to dashboard Cite

Abstract-Multi-rate digital signal processing (DSP) algorithms are often modeled with synchronous dataflow graphs (SDFGs). A lower iteration period implies a faster execution of a DSP algorithm. Retiming is a simple but efficient graph transformation technique for performance optimization, which can decrease the iteration period without affecting functionality. In this paper, we deal with two problems: feasible retimingretiming an SDFG to meet a given iteration period constraint; and optimal retiming -retiming an SDFG to achieve the smallest iteration period. We present a novel algorithm for feasible retiming and based on that one, a new algorithm for optimal retiming, and prove their correctness. Both methods work directly on SDFGs, without explicitly converting them to their equivalent homogeneous SDFGs (HSDFGs). Experimental results show that our methods give a significant improvement compared to the earlier methods.

show abstract

Section: Basic Idea Of Our Methodsmentioning

confidence: 99%

“…Following [15], potentially optimal IPs are checked decreasingly until we find one that is not feasible.…”

Section: Basic Idea Of Our Methodsmentioning

confidence: 99%

Section: Optimal Retimingmentioning

confidence: 99%

mentioning

confidence: 99%

mentioning

confidence: 99%

See 3 more Smart Citations

Efficient Retiming of Multirate DSP Algorithms

Zhu

Basten

Geilen

et al. 2012

IEEE Trans. Comput.-Aided Des. Integr. Circuits Syst.

View full text Add to dashboard Cite

show abstract

Synthesis of Embedded Software

Shukla¹,

Talpin²

2010

View full text Add to dashboard Cite

A Fast Heuristic to Pipeline SDF Graphs

Honorat

Desnos

Dardaillon

et al. 2020

Lecture Notes in Computer Science

View full text Add to dashboard Cite

A common optimization of signal and image processing applications is the pipelining on multiple Processing Elements (PE) available on multicore or manycore architectures. Pipelining an application often increases the throughput at the price of a higher memory footprint. Evaluating different pipeline configurations to select the best one is time consuming: for some applications, there are billions of different possible pipelines. This paper proposes a fast heuristic to pipeline signal and image processing applications modelled with the Synchronous DataFlow (SDF) Model of Computation (MoC). The heuristic automatically adds pipeline stages in the SDF graph in the form of delays, given the Execution Time (ET) of the actors and the number of PEs. The heuristic decreases the time spent by the developer to pipeline its application from hours to seconds. The efficiency of the approach is illustrated with the optimisation of a set of signal and image processing applications executed on multiple PEs. On average, when adding one pipeline stage, our heuristic selects a stage resulting in a better throughput than 90% of all possible stage emplacements.

show abstract

Retiming for Synchronous Data Flow Graphs

Cited by 13 publications

References 7 publications

Efficient Retiming of Multirate DSP Algorithms

Efficient Retiming of Multirate DSP Algorithms

Synthesis of Embedded Software

A Fast Heuristic to Pipeline SDF Graphs

Contact Info

Product

Resources

About