Skewless optical data-link subsystem for massively parallel processors using 8 Gb/s x 1.1 Gb/s MMF array optical module

Yoshikawa, Takeshi; Araki, Satoshi; Miyoshi, Kazunori; Suemura, Y.; Henmi, N.; Nagahori, T.; Matsuoka, Hiroshi; Yokota, Takashi

doi:10.1109/68.643291

Cited by 9 publications

(2 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A hybrid solution is to skip the separate clock channel and encode clock information on the data channels, but still send in bit-parallel mode as reported in Ref. 17. In this case, a deskew unit relying on first in, first out ͑FIFO͒ registers ensures that parallel data words that are output from the receiver are identical with those that were sent.…”

Section: Implementation Aspectsmentioning

confidence: 99%

Two fiber-ribbon ring networks for parallel and distributed computing systems

Jönsson

1998

Opt. Eng

View full text Add to dashboard Cite

Abstract. Ring networks made of fiber-ribbon point-to-point links are proposed. The first network is a control-channel based network in which one fiber in each link joins with others to form a control-channel ring. This ring improves performance of the network by sending medium access control information immediately before the data transmissions. High throughputs can be achieved in the network due to pipelining, i.e., several packets can travel through the network simultaneously but in different segments of the ring. The network can meet tough performance demands in, e.g., massively parallel signal processing systems, which is shown by example. Also, real-time demands can be met using slot reserving. The network, called CC-FPR (control-channel based fiberribbon pipeline ring), can be built today using off-the-shelf fiber optic components. The increasingly good price/performance ratio for fiberribbon links indicates a high potential for the success of the proposed kind of networks; a prototype is currently under development. The second network is similar to first except that it divides the network into two subnetworks, one for packet-switched traffic and one for circuit-switched traffic. When the main data flow in the network does not change rapidly, this is a good choice for a simple but powerful network.

show abstract

Section: Implementation Aspectsmentioning

confidence: 99%

Two fiber-ribbon ring networks for parallel and distributed computing systems

Jönsson

1998

Opt. Eng

View full text Add to dashboard Cite

show abstract

“…The disadvantage is increased hardware complexity when a clock recovery circuit and a buffer circuit for each channel in the receiver are added. A hybrid solution is to skip the separate clock channel and encode clock information on the data channels while still sending in bit-parallel mode, as reported in [24]. In this case, a deskew unit relying on FIFO registers (First In First Out) ensures that parallel data words that are output from the receiver are identical to those which were sent.…”

Section: Fiber-ribbon Linksmentioning

confidence: 99%