Publications

A longstanding performance challenge in datacenter-based applications is how to efficiently handle incoming client requests that spawn many very short (μs scale) jobs that must be handled with high throughput and low tail latency. When no assumptions are made about the duration of individual jobs, or even about the distribution of their durations, this requires blind scheduling with frequent and efficient preemption, which is not scalably supported for μs-level tasks. We present Tiny Quanta (TQ), a system that enables efficient blind scheduling of μs-level workloads. TQ performs fine-grained preemptive scheduling and does so with high performance via a novel combination of two mechanisms: forced multitasking and two-level scheduling. Evaluations with a wide variety of μs-level workloads show that TQ achieves low tail latency while sustaining 1.2x to 6.8x the throughput of prior blind scheduling systems.

Abstract Laboratory evolution is a powerful approach to search for genetic adaptations to new or improved phenotypes, yet either relies on labour-intensive human-guided iterative rounds of mutagenesis and selection, or prolonged adaptation regimes based on naturally evolving cell populations. Here we present CRISPR- and RNA-assisted in vivo directed evolution (CRAIDE) of genomic loci using evolving chimeric donor gRNAs continuously delivered from an error-prone T7 RNA polymerase, and directly introduced as RNA repair donors into genomic targets under either Cas9 or dCas9 guidance. We validate CRAIDE by evolving novel functional variants of an auxotrophic marker gene, and by conferring resistance to a toxic amino acid analogue in baker's yeast Saccharomyces cerevisiae with a mutation rate >3,000-fold higher compared to spontaneous native rate, thus enabling the first demonstrations of in vivo delivery and information transfer from long evolving RNA donor templates into genomic context without the use of in vitro supplied and pre-programmed repair donors.

High-performance, fast-growing natural materials with sustainable and functional features currently arouse significant attention. Here, facile processing, involving delignification, in situ hydrothermal synthesis of TiO2 and pressure densification, is employed to transform natural bamboo into a high-performance structural material. The resulting TiO2-decorated densified bamboo exhibits high flexural strength and elastic stiffness, with both properties more than double that of natural bamboo. Real-time acoustic emission reveals the key role of the TiO2 nanoparticles in enhancing the flexural properties. The introduction of nanoscale TiO2 is found to markedly increase the degree of oxidation and the formation of hydrogen bonds in bamboo materials, leading to extensive interfacial failure between the microfibers, a micro-fibrillation process that results in substantial energy consumption and high fracture resistance. This work furthers the strategy of the synthetic reinforcement of fast-growing natural materials, which could lead to the expanded applications of sustainable materials for high-performance structural applications.

Electrometers measure electric charge, but there must be a fundamental speed limit to measuring one electric charge. Since there are no dimensional inputs to this question, the answer must be expressible in terms of the fundamental physical constants of Nature, e,h,m,c. In general the question should be posed without reference to any specific technology, but for definiteness, we analyze the field effect transistor, which is essentially an electrometer. In spite of selecting a specific technology, we find that the speed limit is related to a fundamental constant, the Rydberg frequency, or as appropriate, the semiconductor Rydberg frequency including the electron effective mass, and the relative dielectric constant. We do not know whether the Rydberg frequency represents the upper speed limit, but on dimensional grounds we claim that the final limit can only differ by some power of the fine-structure-constant.

The interface of planar TiO(2)/polymer photovoltaic cells was modified with two carboxylated polythiophenes having different densities of carboxylic acid groups. Both of the interface modifiers increase the photocurrent of the cells but lower the open-circuit voltage. The work function of the TiO(2), measured using a Kelvin probe, increases with increasing density of carboxylic acid groups due to the formation of interfacial dipoles pointing toward the TiO(2) surface. The formation of interfacial dipoles results in a shift in the band offset at the TiO(2)/polymer interface, which explains the decrease in the open-circuit voltage. This work demonstrates that care must be taken when using carboxylic acid side groups to attach polymers to titania surfaces in photovoltaic cells. If the density of attachment groups is just enough to attach the polymer, then the benefits of the interface modifier can be realized without substantially decreasing the open-circuit voltage.

In their comment, Bui et al . argue that the approach we described in our report is vastly inferior in efficiency to alternative off-the-shelf technologies. Their conclusion is invalid, as they compare efficiencies in completely different operating conditions. Here, using heat transfer and thermodynamics principles, we show how Bui et al .’s conclusions about the efficiencies of off-the-shelf technologies are fundamentally flawed and inaccurate for the operating conditions described in our study.