Proceedings of the ACM/SIGDA International Symposium on Field Programmable Gate Arrays 2012
DOI: 10.1145/2145694.2145705
|View full text |Cite
|
Sign up to set email alerts
|

A mixed precision Monte Carlo methodology for reconfigurable accelerator systems

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

2
28
0

Year Published

2015
2015
2017
2017

Publication Types

Select...
5
1
1

Relationship

1
6

Authors

Journals

citations
Cited by 36 publications
(30 citation statements)
references
References 15 publications
2
28
0
Order By: Relevance
“…8 significant bits and 8 bits for the exponent). As the results indicate, for CP-MCMC simulations where the reduced precision datapaths are used for all the likelihood terms and models the existing hardware design approaches, the mean value of the parameter has a significant bias and also a larger variance compared to the DP-MCMC results, supporting the results obtained in [16], [17]. However, the CF-MCMC sample distributions have the same variance and mean as DP-MCMC samples, which demonstrates that the proposed algorithm removes any bias introduced in the results due to low-precision computations.…”
Section: Quality Of Mcmc Samplessupporting
confidence: 67%
See 3 more Smart Citations
“…8 significant bits and 8 bits for the exponent). As the results indicate, for CP-MCMC simulations where the reduced precision datapaths are used for all the likelihood terms and models the existing hardware design approaches, the mean value of the parameter has a significant bias and also a larger variance compared to the DP-MCMC results, supporting the results obtained in [16], [17]. However, the CF-MCMC sample distributions have the same variance and mean as DP-MCMC samples, which demonstrates that the proposed algorithm removes any bias introduced in the results due to low-precision computations.…”
Section: Quality Of Mcmc Samplessupporting
confidence: 67%
“…Previous works on FPGAs using custom precision can be found in [16], [17], [22], [23]. [23] propose the use of custom precision arithmetic for population-based MCMC methods where multiple parallel chains are used to improve the mixing properties of the chain.…”
Section: Specialised Hardware Approachesmentioning
confidence: 99%
See 2 more Smart Citations
“…5). There are certain improvements that can only be made on specific accelerators, for example utilising multiple pipelines or mixing fixed and floating point calculations on FPGAs [6].…”
Section: E Optimisation Phasesmentioning
confidence: 99%