Qingsheng Hu

Learn More
High-performance input queued switches achieve good performance with low cost. However, with the appearance of optical techniques, the line rate is much higher than before. Scheduling algorithms require not only good performance in delay and stability but fast speed and simple implementation as well. A variety of scheduling algorithms for Virtual Output(More)
Abstracl-This paper presents a novel scalable scheduling architecture for high-performance crossbar-based switches with virtual output queuing (VOQ) scheme. In contrast to traditional switching architectures where the scheduler is implemented by one single centralized scheduling device, the proposed scheduling architecture connects several small scheduling(More)
This paper describes a fine resolution, good linearity, and high throughput time-to-digital converter (TDC), which has realized in 0.18um CMOS technology. Based on a two-channel Vernier delay line (VDL) structure and an asynchronous pipelined readout circuitry, the TDC can achieve a maximum throughput of 500MS/s, a time resolution of 10ps and a total(More)
This paper presents a 6.25Gb/s adaptive decision feedback equalizer (DFE) for high speed backplane communication, which consists of a high-speed DFE and an adaptation engine based on Sign-Sign LMS algorithm. Half-rate structure and current mode logic are employed to pursue a high operating speed. The adaptation engine is composed of sense amplifiers (SA), a(More)
A 6.25Gb/s 3-tap T/2-spaced feed-forward equaliser (FFE) is realized in 0.18μm CMOS Technology. The proposed FFE can be used to reduce the inter-symbol interference (ISI). A high frequency boost delay element using source capacitive degeneration is adopted to meet the high speed requirement. Additionally, a delay locked loop and a load calibration(More)
This paper presents a 10Gb/s concatenated BCH encoder of super FEC. The concatenated BCH encoder module contains 32 BCH column encoder groups and 2 BCH row encoder groups. To achieve high speed, parallel encoding technique is employed. In addition, some optimization methods such as sub-expression sharing and tree-based structure are also adopted to balance(More)