Related papers: Performance Considerations for Gigabyte per Second…
We demonstrate up to 12 km, 56 Gb/s DMT transmission using high-speed VCSELs in the 1.5 um wavelength range for future 400Gb/s intra-data center connects, enabled by vestigial sideband filtering of the transmit signal.
As the needs of Internet users and applications significantly changed over the last decade, inter-domain routing became more important to fulfill these needs. The ways how data flows over the Internet are still completely in the hand of…
High level programming languages and GPU accelerators are powerful enablers for a wide range of applications. Achieving scalable vertical (within a compute node), horizontal (across compute nodes), and temporal (over different generations…
We review some currently discussed solutions for 400 Gbit/s inter-data center WDM transmission for up to 100 km. We focus on direct detected solutions, namely PAM4 and DMT, and present two WDM systems based on these formats.
We experimentally evaluate a flexible DMT system using 4 to 8 50-GHz-grid C-band channels to transmit 448 Gb/s over up to 240 km SSMF. VSB filtering enabled by detuned lasers significantly reduces the impact of chromatic dispersion.
The Large Synoptic Survey Telescope (LSST) is a proposed 8.4-meter telescope that will be located in the Andes mountains in Chile. Every 17 seconds, a 6.4 GB image is transferred to Illinois for immediate processing. That transfer needs to…
Intensive computation is entering data centers with multiple workloads of deep learning. To balance the compute efficiency, performance, and total cost of ownership (TCO), the use of a field-programmable gate array (FPGA) with…
Unprecedented transport efficiency is demonstrated for electrons on the surface of micron-scale superfluid helium filled channels by co-opting silicon processing technology to construct the equivalent of a charge-coupled device (CCD).…
The optimization of the transpose convolution layer for deep learning applications is achieved with the kernel segregation mechanism. However, kernel segregation has disadvantages, such as computing extra elements to obtain the output…
In recent years, emerging storage hardware technologies have focused on divergent goals: better performance or lower cost-per-bit. Correspondingly, data systems that employ these technologies are typically optimized either to be fast (but…
GPUs are broadly used in I/O-intensive big data applications. Prior works demonstrate the benefits of using GPU-side file system layer, GPUfs, to improve the GPU performance and programmability in such workloads. However, GPUfs fails to…
The Domain Name System (DNS) plays a critical role in the functioning of the Internet. It provides a hierarchical name space for locating resources. Data is typically stored in plain text files, possibly spanning gigabytes. Frequent parsing…
Compression algorithms reduce the redundancy in data representation to decrease the storage required for that data. Data compression offers an attractive approach to reducing communication costs by using available bandwidth effectively.…
Transformers have revolutionized AI in natural language processing and computer vision, but their large computation and memory demands pose major challenges for hardware acceleration. In practice, end-to-end throughput is often limited by…
The data production farm for the CDF experiment is designed and constructed to meet the needs of the Run II data collection at a maximum rate of 20 MByte/sec during the run. The system is composed of a large cluster of personal computers…
As a new approach to realizing high-precision time synchronization between remote time scales, quantum two-way time transfer via laboratory fiber link has shown significant enhancement of the transfer stability to several tens of…
With present-day detectors in high energy physics one often faces fast analog pulses of a few nanoseconds length which cover large dynamic ranges. In many experiments both amplitude and timing information have to be measured with high…
Data breaches and cyberattacks represent a severe problem in higher education institutions and universities that can result in illegal access to sensitive information and data loss. To enhance the security of data transmission, Intrusion…
BiSeNet has been proved to be a popular two-stream network for real-time segmentation. However, its principle of adding an extra path to encode spatial information is time-consuming, and the backbones borrowed from pretrained tasks, e.g.,…
We summarise how a 3.0 GHz 16 core AVX512 computer can interpret the equivalent of up to on average 1103370000000 GPop/s. Citations to existing publications are given. Implementation stress is placed on both parallel computing, bandwidth…