Search

This paper presents a complete 200Gb/s PAM-4 transmitter (TX) in 28nm CMOS technology. The transmitter features a hybrid sub-sampling PLL (SSPLL) with a delta-sigma (?S) modulator, clock distribution network with flexible timing control, and data path with a hybrid 5-tap Feed-Forward Equalizer (FFE) and T-coil for bandwidth extension. The prototype chip achieves 4.69 pJ/bit efficiency, 54mV eye height, 0.27UI eye width, and 97% RLM under -6dB channel loss at 50GHz.

This work presents a 16mm2 heterogeneous RISC-V system-on-a-chip (SoC) composed of a high-performance out-of-order core, energy-efficient in-order core, data-parallel vector accelerator, and systolic array deep neural network (DNN) accelerator in low-power Intel 22FFL for general-purpose compute, DNN, and vector workloads. The heterogeneous RISC-V SoC is composed of fully open-source components, including a second-generation Berkeley Out-of-Order Machine (BOOM) with a non-speculative mode attached to a Hwacha vector accelerator, a Rocket in-order core attached to a Gemmini systolic array DNN accelerator, as well as a IMiB L2 cache and off-chip I/Os. Combined, the variety of heterogeneous compute allows for wide programmability while providing up to a 286x MOPS/W improvement or 282x MOPS improvement over the RISC-V in-order core.

The ever-expanding demand for ultra-high-speed interconnects has driven the development of wireline TXs operating at >100Gb/s per lane [1]–[4]. This paper presents a PAM-4 TX achieving 200Gb/s with improved output bandwidth and output swing by minimizing the driver capacitance with pull-up current sources, multiplexing with flexible clock timing control, and employing a fully reconfigurable 5-tap FFE architecture.

Due to I/O bandwidth limitations, intelligent in situ data reduction methods are needed to enable post-hoc workflows. Current state-of-the-art sampling methods save data points if they deem them spatially or temporally important. By analyzing the properties of the data values at each time-step, two consecutive steps may be very similar. This research follows the notion that if neighboring time-steps are very similar, samples from both are unnecessary, which leaves storage for adding more useful samples. Here, we present an investigation of the combination of spatial and temporal sampling to drastically reduce data size without the loss of valuable information. We demonstrate that, by reusing samples, our reconstructed data set reduces the overall data size while achieving a higher post-reconstruction quality over other reduction methods.

As the computation power of supercomputers increases, so does simulation size, which in turn produces orders-of-magnitude more data. Because generated data often exceed the simulation’s disk quota, many simulations would stand to benefit from data-reduction techniques to reduce storage requirements. Such techniques include autoencoders, data compression algorithms, and sampling. Lossy compression techniques can significantly reduce data size, but such techniques come at the expense of losing information that could result in incorrect post hoc analysis results. To help scientists determine the best compression they can get while keeping their analyses accurate, we have developed Foresight, an analysis framework that enables users to evaluate how different data-reduction techniques will impact their analyses. We use particle data from a cosmology simulation, turbulence data from Direct Numerical Simulation, and asteroid impact data from xRage to demonstrate how Foresight can help scientists determine the best data-reduction technique for their simulations.

In recent years, the increased demand for wireless multimedia services has led to huge bandwidth crisis. The concept of Cognitive Radio Networks (CRN) has been proposed as a solution. In CRN, higher Primary User Activity leads to an increased number of Pre-empted Secondary Users (PSU). Since the PSUs are given priority over New Secondary Users (NSU), the NSUs might get starved. Due to this priority of PSUs less number low priority users can be served. The advent of the Internet of Things, the traffic pattern has become unpredictable. All this reasons required a Dynamic Cognitive Channel Allocation (DCCA) algorithm which has been proposed in this work which balances the service among the PSUs and NSUs as per priority of SU as well as avoiding starvation. Analysis of the proposed algorithm using Continuous Time Markov Chain Model shows an improvement over existing literature of 44.44% in NSU dropping probability and 22.8% improvement in NSU blocking probability with respect to arrival of PUs. Varying the arrival of SUs has resulted in an improvement of 58% and 45.75% in the dropping and blocking probabilities of NSUs respectively. The starvation of SUs has lowered by 32.52% with varying arrival rate of PU and 41.38% with varying arrival rate of SUs

Visualization of high-fidelity scientific simulations with high-dimensional inputs and outputs is an important task. Existing high-dimensional data visualization approaches generally assume a substantial amount of data are available or can be generated as needed. However, many of these simulations can be very computationally intensive, taking minutes or hours to run. Analysis and visualization of such expensive simulations poses a challenge. Statistical emulators are frequently used to approximate simulations for statistical analyses. In this work, we propose a visualization tool for an emulator of the simulator and describe how emulators can be used to create effective visualization systems. We choose Gaussian process emulators for this purpose as they enable fast and accurate prediction with uncertainty information. Using these predictions, we design a system that enables visualization of high-dimensional input and output spaces of complex physics simulations. Users of our system can get a detailed understanding of the uncertainties associated with the emulator predictions in both input and output space for a high-dimensional simulation.

Although vortex analysis and detection have been extensively investigated in the past, none of the existing techniques are able to provide fully robust and reliable identification results. Local vortex detection methods are popular as they are efficient and easy to implement, and produce binary outputs based on a user-specified, hard threshold. However, vortices are global features, which present challenges for local detectors. On the other hand, global detectors are computationally intensive and require considerable user input. In this work, we propose a consensus-based uncertainty model and introduce spatial proximity to enhance vortex detection results obtained using point-based methods. We use four existing local vortex detectors and convert their outputs into fuzzy possibility values using a sigmoid-based soft-thresholding approach. We apply a majority voting scheme that enables us to identify candidate vortex regions with a higher degree of confidence. Then, we introduce spatial proximity- based analysis to discern the final vortical regions. Thus, by using spatial proximity coupled with fuzzy inputs, we propose a novel uncertainty analysis approach for vortex detection. We use expert's input to better estimate the system parameters and results from two real-world data sets demonstrate the efficacy of our method.

When the spatial and temporal resolutions of a time-varying simulation become very high, it is not possible to process or store data from every time step due to the high computation and storage cost. Although using uniformly down-sampled data for visualization is a common practice, important information in the un-stored data can be lost. Currently, linear interpolation is a popular method used to approximate data between the stored time steps. For pathline computation, however, errors from the interpolated velocity in the time dimension can accumulate quickly and make the trajectories rather unreliable. To inform the scientist the error involved in the visualization, it is important to quantify and display the uncertainty, and more importantly, to reduce the error whenever possible. In this paper, we present an algorithm to model temporal interpolation error, and an error reduction scheme to improve the data accuracy for temporally down-sampled data. We show that it is possible to compute polynomial regression and measure the interpolation errors incrementally with one sequential scan of the time-varying flow field. We also show empirically that when the data sequence is fitted with least-squares regression, the errors can be approximated with a Gaussian distribution. With the end positions of particle traces stored, we show that our error modeling scheme can better estimate the intermediate particle trajectories between the stored time steps based on a maximum likelihood method that utilizes forward and backward particle traces.