Rapid system prototyping for high performance reconfigurable computing

Shrivastava, Sarvesh; Jain, V.K.

doi:10.1109/iwrsp.1999.779028

Cited by 4 publications

(6 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The MA_PLUS cell was described in [8], [16], [17] therefore we will not discuss it here. Suffice it to say, that it provides flexibility in computation by use of a reconfiguration register and a rich mix of multiply-add, multiply-accumulate, absolute difference, max/min, and muxing functions.…”

Section: Coarse Grain Cellsmentioning

confidence: 99%

“…Then, for the computation in (8), the polyphase filter h (i) (m) must be either accessed from a ROM, or computed on-line for software radio application.…”

Section: Implementation On J-platformmentioning

confidence: 99%

“…Rate Conversion for Software Radio: 0.25 micron CMOS From Section 5, we know that a JIF module needs 8 UNLs, 8 MA_PLUSs, and 4 DFs. A single module is adequate for this rate conversion function for L=8 case, and six are needed for the L=48 case.…”

Section: Performance Estimatesmentioning

confidence: 99%

“…Our recent work has demonstrated that high performance computing is possible on the applications oriented Jplatform [8], [16], [17].…”

Section: Introductionmentioning

confidence: 99%

“…The MA_PLUS, which is an enhanced multiply-add cell, and the UNL, a Universal NonLinear cell. The MA_PLUS cell can perform any of the functions such as multiply-add, multiply-accumulate, two separate multiply and add operations simultaneously, absolute difference, or the min/max operation, among others [8]. The UNL cell, which we have devised and refined over several years [3]- [6], can perform any one of several nonlinear operations (such as the familiar trigonometric functions, inverse operations, square-root, inverse-squareroot, etc.)…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Mapping a high-speed wireless communication function to the reconfigurable J-platform

Jain

Proceedings 11th International Workshop on Rapid System Prototyping. RSP 2000. Shortening the Path From Specification to Protot

Self Cite

View full text Add to dashboard Cite

This paper maps a software radio function to a coarse grain platform, namely the J-Platform. In our previous work the J-platform was based on two types of cells, namely the MA_PLUS, which is an enhanced multiply-add cell, and the UNL, a Universal NonLinear cell. Here, we introduce one more cell, namely the DF cell, which is a Data Fabric cell. These three cells account for the versatility of the approach and provide an unprecedented capability for 'coarse-grain reconfigurable computing'. Although the applications range from FIR filtering of images and video to large-scale inverse problems, and from reconstruction of computerized tomography images to color conversion of Video from the RGB to HSI domain, here we present a new type of application, namely a computation intensive function of software radio. Specifically, we discuss very high 'spur-free dynamic range (SFDR)' interpolation for sample rate conversion in IF band. The high SFDR values --up to 130 dB --are achieved through a new family of interpolation filters as proposed in the paper. Significant improvement in both performance and design time is thereby achieved.

show abstract

Section: Coarse Grain Cellsmentioning

confidence: 99%

“…Then, for the computation in (8), the polyphase filter h (i) (m) must be either accessed from a ROM, or computed on-line for software radio application.…”

Section: Implementation On J-platformmentioning

confidence: 99%

Section: Performance Estimatesmentioning

confidence: 99%

“…Our recent work has demonstrated that high performance computing is possible on the applications oriented Jplatform [8], [16], [17].…”

Section: Introductionmentioning

confidence: 99%