Yigit Demir

EcoLaser: an adaptive laser control for energy-efficient on-chip photonic interconnects

The high-speed and low-cost modulation of light make photonic interconnects an attractive solution for the communication demands of manycore processors. However, the high optical loss of many nanophotonic components results in high laser power consumption, most of which is wasted during periods of system inactivity. We propose EcoLaser, an adaptive laser control mechanism that saves between 24-77% of the laser power by turning off the laser when not needed. These power savings allow the cores to exploit a higher power budget and achieve speedups of 1.1-2x.

SLaC: Stage laser control for a flattened butterfly network

Photonic interconnects have emerged as a promising candidate technology for high-performance energy-efficient on-chip, on-board, and datacenter-scale interconnects. However, the high optical loss of many nanophotonic components coupled with the low efficiency of current laser sources result in exceedingly high total power requirements for the laser. As optical interconnects stay on even during periods of system inactivity, most of this power is wasted, which has prompted research on laser gating. Unfortunately, prior work on laser gating has only focused on low-scalability on-chip photonic interconnects (photonic crossbars), and disrupts the connectivity of the network which renders a high-performance implementation challenging. In this paper we propose SLaC, a laser gating technique that turns on and off redundant paths in a photonic flattened-butterfly network to save laser energy while maintaining high performance and full connectivity. Maintaining full connectivity removes the laser turn-on latency from the critical path and results in minimal performance degradation. SLaC is equally applicable to on-chip, on-board, and datacenter level interconnects. For on-chip and multi-chip applications, SLaC saves up to 67% of the laser energy (43-57% on average) when running real-world workloads. On a datacenter network, SLaC saves 79% of the laser energy on average when running traffic traces collected from university datacenter servers.

Parka: Thermally Insulated Nanophotonic Interconnects

Silicon-photonics are emerging as the prime candidate technology for energy-efficient on-chip interconnects at future process nodes. However, current designs are primarily based on microrings, which are highly sensitive to temperature. As a result, current silicon-photonic interconnect designs expend a significant amount of energy heating the microrings to a designated narrow temperature range, only to have the majority of the thermal energy waste away and dissipate through the heat sink, and in the process of doing so heat up the logic layer, causing significant performance degradation to the cores and inducing thermal emergencies. We propose Parka, a nanophotonic interconnect that encases the photonic die in a thermal insulator that keeps its temperature stable with low energy expenditure, while minimizing the spatial and temporal thermal coupling between logic and silicon-photonic components. Parka reduces the microring energy by 3.8--5.4x and achieves 11--23% speedup on average (34% max) depending on the cooling solution used.

Towards energy-efficient photonic interconnects

Silicon photonics have emerged as a promising solution to meet the growing demand for high-bandwidth, low-latency, and energy-efficient on-chip and off-chip communication in many-core processors. However, current silicon-photonic interconnect designs for many-core processors waste a significant amount of power because (a) lasers are always on, even during periods of interconnect inactivity, and (b) microring resonators employ heaters which consume a significant amount of power just to overcome thermal variations and maintain communication on the photonic links, especially in a 3D-stacked design. The problem of high laser power consumption is particularly important as lasers typically have very low energy efficiency, and photonic interconnects often remain underutilized both in scientific computing (compute-intensive execution phases underutilize the interconnect), and in server computing (servers in Google-scale datacenters have a typical utilization of less than 30%). We address the high laser power consumption by proposing EcoLaser+, which is a laser control scheme that saves energy by predicting the interconnect activity and opportunistically turning the on-chip laser off when possible, and also by scaling the width of the communication link based on a runtime prediction of the expected message length. Our laser control scheme can save up to 62 - 92% of the laser energy, and improve the energy efficiency of a manycore processor with negligible performance penalty. We address the high trimming (heating) power consumption of the microrings by proposing insulation methods that reduce the impact of localized heating induced by highly-active components on the 3D-stacked logic die.

Harnessing path diversity for laser control in data center optical networks

Optical interconnects are already the dominant technology in large-scale datacenter networks. Unfortunately, the high optical loss of many optical components, coupled with the low efficiency of laser sources, result in high aggregate power requirements for the thousands of optical transceivers that such networks employ. As optical interconnects stay always on, even during periods of system inactivity, most of this power is wasted. Ideally we would like to turn off the transceivers when a network link is idle (i.e., “power gate” the lasers), and turn them back on right before the next transmission. The danger with this approach is that it may expose the laser turn-on delay and lead to higher network latency. However, data center networks typically employ network topologies with path diversity and facilitate multiple paths for each source-destination pair. Based on this observation, we propose an optical network architecture where redundant paths are turned off when the extra bandwidth they provide is not needed, and they turn back on when traffic increases beyond a high watermark to decongest the network. Maintaining full connectivity removes the laser turn-on latency from the critical path and results in minimal performance degradation, while at the same time power-gating the lasers saves 60% of the laser power on average on a variety of data center traffic scenarios.

Energy-Proportional Photonic Interconnects

Photonic interconnects have emerged as the prime candidate technology for efficient networks on chip at future process nodes. However, the high optical loss of many nanophotonic components coupled with the low efficiency of current laser sources results in exceedingly high total power requirements for the laser. As optical interconnects stay on even during periods of system inactivity, most of this power is wasted, which has prompted research on laser gating. Unfortunately, prior work has been complicated by the long laser turn-on delays and has failed to deliver the full savings. In this article, we propose ProLaser, a laser control mechanism that monitors the requests sent on the interconnect, the cache, and the coherence directory to detect highly correlated events and turn on proactively the lasers of a photonic interconnect. While ProLaser requires fast lasers with a turn-on delay of a few nanoseconds, a technology that is still experimental, several types of such lasers that are suitable for power gating have already been manufactured over the last decade. Overall, ProLaser saves 42% to 85% of the laser power, outperforms the current state of the art by 2× on average, and closely tracks (within 2%--6%) a perfect prediction scheme with full knowledge of future interconnect requests. Moreover, the power savings of ProLaser allow the cores to exploit a higher-power budget and run faster, achieving speedups of 1.5 to 1.7× (1.6× on average).