"Ayar Labs: Faster I/O" BY: KEVIN FOGARTY in Semiconductor Engineering

Startup AyarLabs is using a combination of high-bandwidth fiberoptics, low-cost CMOS fabrication and careful target selection to strike efficiently at the datacenter’s worst bottleneck.

“Moore’s Law only covers the processor, not how we move data in and out of it during processing or how to get the processor and memory working at the same speed,” according to Alexandra Wright-Gladstein, co-founder and CEO of Ayar Labs. “We are developing optical transceivers, but we’re very focused on chip I/O , not optics per se. We have a high-efficiency optical-electrical conversion engine with a very small footprint, so our optical devices are tiny. We build it using ordinary CMOS design and fabrication and focus on efficiency, so we end up with a product that uses 10 times less power than any commercial optical transceiver, whether silicon-based or traditional.”

The company is a spin-out from a 10-year MIT research project that developed a dual-core RISC-V architecture chip with 70 million transistors,850 million photonic components, 1MB of RAM, and a fiber I/O port theoretically capable of sending at at 550 Gb/sec and receiving 900Gb/sec.

When the team combined two chips, one serving a processor, the other as memory, they were able to pass data at 2.5Gbit/sec in each direction, expanding bandwidth as needed by adding wavelengths from the external laser that provided light to the chips.

The project was partly funded by two DARPA programs, one designed to increase data rates by combining processing with photonics, and another whose focus was on getting the highest number of FLOPS possible from every watt.

A paper describing the result was published in the journal Nature in December, 2015, complete with a video showing how it works.

The SoC approach is designed partly for efficiency and partly to bridge the gap in design and manufacturing between CMOS and photonics products, according to the paper, which also describes the “zero-change” approach to developing photonics using CMOS techniqes and materials in the integration of photonics to avoid the complication of custom fabrication of photonic components.

The 45nm chip was the first designed with an optical I/O. The commercial product is a lot different than the one described in the journal article, according to Wright-Gladstein, but the research version did more than just give co-founders Chen Sun and Mark Wade a chance to pursue their Ph.D.s.

It also brought in some startup capital after Wright-Gladstein entered the nascent company — then called Optibit — in MIT’s Clean Energy Prize competition and brought home both grand prizes, which totaled $275,000. https://www.scientificcomputing.com/news/2015/05/optical-chips-team-develops-way-integrate-fiber-optics-computer-chips

Optics give datacenters tools to retool
Intel’s 100 Gbit/sec silicon photonic network interface gave datacenter architects the opportunity to change the fundamental design of datacenters by eliminating much of the penalty of distance. Minimizing the delay in providing a Google Translation or passing on a Facebook “Like” used to mean larding each individual server with as much memory as possible, and keeping HDD storage in the rack or very close by so that data not stored in main memory didn’t have far to go.

Replacing copper running at 10Gbit/sec, 25 Gbit/sec or even 40Gbitsec with fiber running 100Gbit/sec, while using less energy and generating less heat, allows hyperscale datacenters to pack in even more hardware and delay just a little the need to build a new one. The year it delivered its first photonics product, 2016, was also the first time that cloud service providers bought more Intel chips than Intel’s traditional customers, according to a 451 Research paper about the project.

Sales of silicon photonic switches generated only $30 million during 2016, according to researchers at Yole Développement, but will make up 35% of the market for intra-datacenter communication by 2025, generating sales worth $560 million. Cloud service companies and large data centers are doing most of the spending to increase the bandwidth of rack-top switches, and increasingly to link geographically separate facilities using wave-division multiplexing. WDM sales rose 225% during 2017 compared to 2016, according to analysts at the Dell’Oro Group.

Use negative feedback opamp to understand the feedback divider in Phase-locked loop

Phase-locked loop is a feedback system. In the feedback loop, a divider is usually used to get a frequency multiplication.

At the first sight, it is hard to understand why the output frequency feedback with a divider can get a increased frequency clock signal.

I found a article in ALL ABOUT CIRCUITS website. It is very useful to explain it by using the negative feedback opamp knowledge.

In the opamp non-inverting amplifier, if the output signal is connected directly back to the inverting input termal, it is used as a voltage follower, or unity-gain buffer. As a result, the opamp does whatever it needs to do to make the output voltage equal to the input voltage.

What if we need a gain? We just use some resistors to turn the feedback loop into a voltage divider. The feedback voltage, Vfb, will be a division of output voltage, Vout.

As we know, the negative-feedback arrangement causes the op-amp to modify its output with one goal: make the voltage at two input nodes equal. When it is configured as unity-gain buffer, this means the Vout must be equal Vin.

But the voltage divider in the feedback loop changes everything. Now, the voltage at the inverting input is DIV times smaller than the voltage at the output. Thus, in order to make the inverting-input voltage equal to the noninverting-input voltage, the output voltage must be DIV times larger than the input voltage.

With an op-amp, then, we create voltage gain by reducing the amplitude of the feedback voltage; with a PLL, we create frequency gain by reducing the frequency of the feedback waveform. To continue the analogy, the gain of a noninverting opamp circuit is equal to the factor by which the feedback voltage is divided, and the amount of frequency multiplication perfomed by the PLL is equal to the factor by which the feedback signal's frequency is divided.

Fundamentals of Data conversion

I really like a excellent summary about data conversion:

"No matter which ADC and what kind of ADC architecture, they all perform two basic and fundamental operations: discretization in time and discretization in amplitude. "

The first step, time discretization, is also called sampling operation. The continually time-varying input analog signal is sampled at uniformly spaced times at a frequency of fs and the samples are thus separated by a period of T=1/fs.

Although the sampled signal are obtained, they are still has different infinite range of values in amplitude domain. Therefore, we can't be represent them precisely in digital codes.

The second step is to digitize the sampled signal in amplitude. The so-called quantization error limits the resolution of the converter.

There are two broad categories of ADCs: Nyquist-rate and over-sampling converters. They typically offer different compromises between ADC resolution and output sampling rate.

Nyquist-rate converters are those that operate at the minimum sampling frequency necessary to capture all the information about the entire input bandwidth, and therefore the output data rate of a Nyquist-rate ADC can be very high.

Over-sampling converters are sampling at very high frequency. The noise-shaping technique is used to filter the in-band noise and to achieve higher SNDR and SFDR. Two types of over-sampling converters are developed currently, DT and CT. Indeed CT sigma delta ADC is introduced earlier in 1962 than DT sigma delta ADC. But when switched-capacitor circuits were introduced, most CT sigma delta modulators were implemented with DT loop filters. SC circuits remain popular because of their insensitivity to signal waveform characteristics.

In addition, the time constants of SC integrators scale with sampling frequency, allowing for greater system flexibility. However, interest in CT sigma delta modulators has been renewed because of their benefits versus sampled-input ADCs, such as employing lower-power integrator amplifiers and including inherent anti-aliasing filtering.

Because of the over-sampling operation, the output rates of sigma-delta ADCs are currently limited to less than 100MSPS while pipleine ADCs are capable of operating up to 500MSPS and beyond. Indeed, for a given technology, Nyquist-rate converters will always be able to operate faster than sigma-delta ADCs because of the over-sampling necessary in a sigma-delta design. Fortunately, the benefts of CT sigma delta technology outweigh the drawbacks for high-resolution applications at sampling rates below 100MSPS.