A Linear Complexity Algorithm for the Automatic Generation of Convex Multiple Input Multiple Output Instructions

Galuzzi, Carlo; Bertels, Koen; Vassiliadis, S.

doi:10.1007/978-3-540-71431-6_13

Cited by 11 publications

(12 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [3], despite the different benchmarks, we both use cryptographic benchmarks and the overall application speed-up has the same order of magnitude (x5.5 with our approach and x7.5 with the approach presented in [3] but which still has high overall computational complexity in contrast with the linear complexity of our approach. Compared with the work presented in [11] (Figure 5), the clustering algorithms presented in this paper perform better in most of the cases.…”

Section: B Analysis Of the Datamentioning

confidence: 62%

“…Since MIMO instructions are combinations of MISO instructions and the combination of convex instructions is not always convex, we have to provide a way to combine convex instructions guaranteeing the convexity property of the final MIMO instruction generated. In [13], [11] we presented the following results (Theorem III.5 and Corollary III.6):…”

Section: Definition Iii2 a Subgraph G *mentioning

confidence: 86%

“…is a convex cluster, where NIn(m) is the number of inputs of the node m and NIn C (m) is the number of inputs coming from a set C of the node m. The proof is a particular case of the proof in [11].…”

Section: Definition Iv1mentioning

confidence: 99%

“…Additionally, if in place of STEP D1, we reiterate STEP C1 increasing one level at each iteration (i → i + 1) until there are nodes that can be included in the cluster, the algorithm generates convex MIMO clusters as in [11] with or without reiteration of the initial steps (i.e. with or without STEP B1 and B2).…”

Section: Definition Iv1mentioning

confidence: 99%

“…In [11], [10], we have proposed two clustering algorithms (based on the previous Theorem III. 5) which generate convex MIMO instruction-set extensions.…”

Section: Corollary Iii6 Any Combination Of Maxmisos At the Same Levementioning

confidence: 99%

See 4 more Smart Citations

Algorithms for the automatic extension of an instruction-set

Galuzzi¹,

Theodoropoulos²,

Meeuws³

et al. 2009

2009 Design, Automation &Amp; Test in Europe Conference &Amp; Exhibition

View full text Add to dashboard Cite

In this paper, two general algorithms for the automatic generation of instruction-set extensions are presented. The basic instruction set of a reconfigurable architecture is specialized with new applicationspecific instructions. The paper proposes two methods for the generation of convex multiple input multiple output instructions, under hardware resource constraints, based on a two-step clustering process. Initially, the application is partitioned in single-output instructions of variable size and then, selected clusters are combined in convex multiple output clusters following different policies. Our results on well-known kernels show that the extended Instructions-Set allows to execute applications more efficiently and needing fewer cycles. Our results show that a significant overall application speed-up is achieved even for large kernels (for ADPCM decoder the speed-up is up to x2.2 and for TWOFISH encoder the speedup is up to x5.5).

show abstract

Section: B Analysis Of the Datamentioning

confidence: 62%

Section: Definition Iii2 a Subgraph G *mentioning

confidence: 86%

Section: Definition Iv1mentioning

confidence: 99%

Section: Definition Iv1mentioning

confidence: 99%

“…In [11], [10], we have proposed two clustering algorithms (based on the previous Theorem III. 5) which generate convex MIMO instruction-set extensions.…”

Section: Corollary Iii6 Any Combination Of Maxmisos At the Same Levementioning

confidence: 99%

See 3 more Smart Citations