We examine the current performance and future demands of interconnects to and on silicon chips. We compare electrical and optical interconnects and project the requirements for optoelectronic and optical devices if optics is to solve the major problems of interconnects for future high-performance silicon chips. Optics has potential benefits in interconnect density, energy, and timing. The necessity of low interconnect energy imposes low limits especially on the energy of the optical output devices, with a ~ 10 fJ/bit device energy target emerging. Some optical modulators and radical laser approaches may meet this requirement. Low (e.g., a few femtofarads or less) photodetector capacitance is important. Very compact wavelength splitters are essential for connecting the information to fibers. Dense waveguides are necessary on-chip or on boards for guided wave optical approaches, especially if very high clock rates or dense wavelength-division multiplexing (WDM) is to be avoided. Free-space optics potentially can handle the necessary bandwidths even without fast clocks or WDM. With such technology, however, optics may enable the continued scaling of interconnect capacity required by future chips.

Device Requirements for Optical Interconnects to Silicon Chips

This paper describes both a near term and a long term optical interconnect solution, the first based on a packaging architecture and the second based on a monolithic photonic CMOS architecture. The packaging-based optical I/O architecture implemented with 90 nm CMOS transceiver circuits, 1 × 12 VCSEL/detector arrays and polymer waveguides achieves 10 Gb/s/channel at 11 pJ/b. A simple TX pre-emphasis technique enables a potential 18 Gb/s at 9.6 pJ/b link efficiency. Analysis predicts this architecture to reach less than 1 pJ/b at the 16 nm CMOS technology node. A photonic CMOS process enables higher bandwidth and lower energy-per-bit for chip-to-chip optical I/O through integration of electro-optical polymer based modulators, silicon nitride waveguides and polycrystalline germanium (Ge) detectors into a CMOS logic process. Experimental results for the photonic CMOS ring resonator modulators and Ge detectors demonstrate performance above 20 Gb/s and analysis predicts that photonic CMOS will eventually enable energy efficiency better than 0.3 pJ/b with 16 nm CMOS. Optical interconnect technologies such as these using multi-lane communication or wavelength division multiplexing have the potential to achieve TB/s interconnect and enable platforms suitable for the tera-scale computing era.

Optical I/O technology for tera-scale computing

Large capacity memory systems are constructed using stacked memory integrated circuits or chips. The stacked memory chips are constructed in such a way that eliminates problems such as signal integrity while still meeting current and future memory standards.

Methods and apparatus of stacking DRAMs

We present an area-efficient neural signal-acquisition system that uses a digitally intensive architecture to reduce system area and enable operation from a 0.5 V supply. The architecture replaces ac coupling capacitors and analog filters with a dual mixed-signal servo loop, which allows simultaneous digitization of the action and local field potentials. A noise-efficient DAC topology and an compact, boxcar sampling ADC are used to cancel input offset and prevent noise folding while enabling “per-pixel” digitization, alleviating system-level complexity. Implemented in a 65 nm CMOS process, the prototype occupies 0.013 mm2 while consuming 5 μW and achieving 4.9 μVrms of input-referred noise in a 10 kHz bandwidth.

A 0.013 ${\hbox {mm}}^{2}$ , 5 $\mu\hbox{W}$ , DC-Coupled Neural Signal Acquisition IC With 0.5 V Supply

This paper reviews recent progress and future directions of signal integrity design for high-speed digital circuits, focusing on four areas: signal propagation on transmission lines, discontinuity modeling and characterization, measurement techniques, and link-path design and analysis.

Signal Integrity Design for High-Speed Digital Circuits: Progress and Directions

We present a scalable low-power I/O transceiver in 65 nm CMOS, capable of 5-15 Gbps operation over single-board and backplane FR4 channels with power efficiencies between 2.8-6.5 mW/Gbps. Nonlinear power-performance tradeoff is achieved by the use of scalable transceiver circuit blocks and joint optimization of the supply voltage, bias currents and driver power with data rate. Low-power operation is enabled by passive equalization through inductive link termination, active continuous-time RX equalization, global TX/RX clock distribution with on-die transmission lines, and low-noise offset-calibrated receivers.

A Scalable 5–15 Gbps, 14–75 mW Low-Power I/O Transceiver in 65 nm CMOS

A source-synchronous I/O link with adaptive receiver-side equalization has been implemented in 0.13-/spl mu/m bulk CMOS technology. The transceiver is optimized for small area (360 /spl mu/m /spl times/ 360 /spl mu/m) and low power (280 mW). The analog equalizer is implemented as an 8-way interleaved, 4-tap discrete-time linear filter. The equalization improved the data rate of a 102 cm backplane interconnect by 110%. On-die adaptive logic determines optimal receiver settings through comparator offset cancellation, data alignment of the transmitter and receiver, clock de-skew and setting filter coefficients for equalization. The noise-margin degradation due to statistical variation in converged coefficient values was less than 3%.

An 8Gb/s source-synchronous I/O link with adaptive receiver equalization, offset cancellation and clock deskew

Improvements in signaling methods, circuits and process technology have allowed input/output (I/O) data rates to scale beyond 10 Gb/s over several legacy channels. In this regime, it is critical to accurately model and comprehend channel/circuit nonidealities in order to co-optimize the link architecture, circuits, and interconnect. Empirical and worst-case analysis methods used at lower rates are inadequate to account for several deterministic and random noise sources present in I/O links today. In this paper, we review models and methods for statistical signaling analysis of high-speed links, and also propose a new way to integrate behavioral modeling approaches with analytical methods. A computationally efficient segment-based analysis method is shown to accurately capture the effect of transmit jitter and its interaction with the channel. In addition, a new jitter interpretation approach is proposed to enable the analysis of arbitrary I/O clocking topologies. We also present some examples to illustrate the practical utility of these analysis methods in the realm of high-speed I/O design.

Modeling and Analysis of High-Speed I/O Links

A full-duplex transceiver capable of 8-Gb/s data rates is implemented in 0.18-/spl mu/m CMOS. This equalized transceiver has been optimized for small area (329 /spl mu/m /spl times/ 395 /spl mu/m) and low power (158 mW) for point-to-point parallel links. Source-synchronous clocking and per-pin skew compensation eliminate coding bandwidth overhead and reduce latency, jitter, and complexity. This link is self-configuring through the use of automatic comparator offset trim and adaptive deskew. Comprehensive diagnostic capabilities have been integrated into the transceiver to provide link, interconnect, and circuit characterization without the use of external test equipment. With a resolution of 4 mV and 9 ps, these capabilities enable on-die eye diagram generation, equivalent time waveform capture, noise characterization, and jitter distribution measurements.

An 8-Gb/s simultaneous bidirectional link with on-die waveform capture

A memory device having memory cells supplied with a separate higher voltage power than the separate power supplied to memory logic, and a lower power state that entails removing power from at least some of the logic such that refresh operations to preserve the contents of the memory cells continue to take place, but at least some of the interface to the memory device is powered down to reduce power consumption.

Joseph T. Kennedy

Papers

A Scalable 5–15 Gbps, 14–75 mW Low-Power I/O Transceiver in 65 nm CMOS

An 8Gb/s source-synchronous I/O link with adaptive receiver equalization, offset cancellation and clock deskew

Modeling and Analysis of High-Speed I/O Links

An 8-Gb/s simultaneous bidirectional link with on-die waveform capture

Memory system segmented power supply and control