Search for: All records

We present ContextPrefetcher, a host-guided high-performant prefetching framework for near-storage accelerators that prefetches data blocks from storage (e.g., NAND) to devicelevel RAM. Efficiently prefetching data blocks to device-level RAM reduces storage access costs and improves I/O performance. We introduce a novel abstraction, Cross-layered Context (CLC), a virtual entity that spans across the host and the device and is used for identifying, managing, and tracking active and inactive data such as files, objects (within object stores), or a range of blocks. To support efficient prefetching of actively used CLCs to device memory without incurring near-device resource (memory and compute) bottlenecks, ContextPrefetcher delegates prefetching management to the host, guiding near-device compute to prefetch blocks of active CLC. Finally, ContextPrefetcher facilitates the swift reclamation of blocks associated with inactive CLC. Preliminary evaluation against state-of-the-art near-storage accelerator designs demonstrates performance gains of up to 1.34×. 

We introduce CrossPrefetch, a novel cross-layered I/O prefetching mechanism that operates across the OS and a user-level runtime to achieve optimal performance. Existing OS prefetching mechanisms suffer from rigid interfaces that do not provide information to applications on the prefetch effectiveness, suffer from high concurrency bottlenecks, and are inefficient in utilizing available system memory. CrossPrefetch addresses these limitations by dividing responsibilities between the OS and runtime, minimizing overhead, and achieving low cache misses, lock contentions, and higher I/O performance. CrossPrefetch tackles the limitations of rigid OS prefetching interfaces by maintaining and exporting cache state and prefetch effectiveness to user-level runtimes. It also addresses scalability and concurrency bottlenecks by distinguishing between regular I/O and prefetch operations paths and introduces fine-grained prefetch indexing for shared files. Finally, CrossPrefetch designs low-interference access pattern prediction combined with support for adaptive and aggressive techniques to exploit memory capacity and storage bandwidth. Our evaluation of CrossPrefetch, encompassing microbenchmarks, macrobenchmarks, and real-world workloads, illustrates performance gains of up to 1.22x-3.7x in I/O throughput.We also evaluate CrossPrefetch across different file systems and local and remote storage configurations. 

Candidate bacterial phylum CSP1-3 has not been cultivated and is poorly understood. Here, we analyzed 112 CSP1-3 metagenome-assembled genomes and showed they are likely facultative anaerobes, with 3 of 5 families encoding autotrophy through the reductive glycine pathway (RGP), Wood–Ljungdahl pathway (WLP) or Calvin-Benson-Bassham (CBB), with hydrogen or sulfide as electron donors. Chemoautotrophic enrichments from hot spring sediments and fluorescence in situ hybridization revealed enrichment of six CSP1-3 genera, and both transcribed genes and DNA-stable isotope probing were consistent with proposed chemoautotrophic metabolisms. Ancestral state reconstructions showed that the ancestors of phylum CSP1-3 may have been acetogens that were autotrophic via the RGP, whereas the WLP and CBB were acquired by horizontal gene transfer. Our results reveal that CSP1-3 is a widely distributed phylum with the potential to contribute to the cycling of carbon, sulfur and nitrogen. The name Sysuimicrobiota phy. nov. is proposed.

 

We propose OmniCache, a novel caching design for nearstorage accelerators that combines near-storage and host memory capabilities to accelerate I/O and data processing. First, OmniCache introduces a “near-cache” approach, maximizing data access to the nearest cache for I/O and processing operations. Second, OmniCache presents collaborative caching for concurrent I/O and data processing by using host and device caches. Third, OmniCache incorporates a dynamic modeldriven offloading support, which actively monitors hardware and software metrics for efficient processing across host and device processors. Finally, OmniCache explores the extensibility for newly-introduced CXL, a memory expansion technology. OmniCache demonstrates significant performance gains of up to 3.24X for I/O workloads and 3.06X for data processing workloads. 

Abstract. The mid-Pliocene Warm Period (mPWP, 3.3–3.0 Ma) was characterised by an atmospheric CO2 concentration exceeding 400 ppmv with minor changes in continental and orbital configurations. Simulations of this past climate state have improved with newer models but still show some substantial differences from proxy reconstructions. There is little information about atmospheric aerosol concentrations during the Pliocene, but previous work suggests that it could have been quite different from the modern period. Here we apply idealised aerosol scenario experiments to examine the importance of aerosol forcing on mPWP tropical precipitation and the possibility of aerosol uncertainty explaining the mismatch between reconstructions and simulations. The absence of industrial pollutants leads to further warming, especially in the Northern Hemisphere. The Intertropical Convergence Zone (ITCZ) becomes narrower and stronger and shifts northward after removal of anthropogenic aerosols. Though not affecting the location of monsoon domain boundary, removal of anthropogenic aerosol alters the amount of rainfall within the domain, increasing summer rain rate over eastern and southern Asia and western Africa. This work demonstrates that uncertainty in aerosol forcing could be the dominant driver in tropical precipitation changes during the mid-Pliocene: causing larger impacts than the changes in topography and greenhouse gases.