Compute Express Link (CXL) 3.0 can turn stranded DRAM into a composable datacenter memory fabric, but hardware is scarce and cycle-accurate simulation is prohibitively slow for full workloads. CXLMemSim is an open-source software CXL memory simulator that keeps computation at native speed and models only CXL-affected memory behavior. Across SPEC CPU2017, Memcached, wrf, vector search, and llama.cpp, CXLMemSim stays within 11–15% of hardware and gem5 while completing a Redis-scale run in 45 minutes instead of gem5’s 73 hours. The public release at SlugLab/CXLMemSim provides PEBS/LBR/eBPF tracing, JSON topology descriptions, and reusable migration/cache policy hooks for CXL design-space exploration.

In SC20, (Srivastava et al. 2020) proposed a Parallel Framework forBayesianLearning, or ramBLe, for short, which is a highly parallel and efficient framework for learning the structure of Bayesian Networks (BNs) from samples,There was a discrepancy in Bibliography in the PDF and the source file. We have followed the source file. ?> particularly large genome-scale networks. As part of our participation in the SC21 Student Cluster Competition, our task was to verify conclusions from the original work (Srivastava et al. 2020). Here we present the outcome of our experiments, which were performed on a four-node cluster from the Oracle Cloud HPC platform. We reproduce the numerical results from (Srivastava et al. 2020), namely the algorithm's performance and scaling behavior using MPI and different Python and Boost libraries on the Oracle cloud.

Memory-Centric X-ray Computational Tomography(CT) is an iterative reconstruction technique that trades compute simplifications with higher memory accesses. MemXCT implements a sparse matrix-vector multiplication(SpMV) with multi-stage buffering and two-level pseudo-Hilbert ordering for optimization. Motivated by the need to validate conclusions from previous work, we reproduce the numerical results, the algorithm’s performance, and the scaling behavior of the algorithms as the number of MPI processes increases on Azure. Digital artifacts from these experiments are available at: 10.5281/zenodo.5598108

As a new programming paradigm, deep learning has expanded its application to many real-world problems. At the same time, deep learning based software are found to be vulnerable to adversarial attacks. Though various defense mechanisms have been proposed to improve robustness of deep learning software, many of them are ineffective against adaptive attacks. In this work, we propose a novel characterization to distinguish adversarial examples from benign ones based on the observation that adversarial examples are significantly less robust than benign ones. As existing robustness measurement does not scale to large networks, we propose a novel defense framework, named attack as defense (A2D), to detect adversarial examples by effectively evaluating an example’s robustness. A2D uses the cost of attacking an input for robustness evaluation and identifies those less robust examples as adversarial since less robust examples are easier to attack. Extensive experiment results on MNIST, CIFAR10 and ImageNet show that A2D is more effective than recent promising approaches. We also evaluate our defense against potential adaptive attacks and show that A2D is effective in defending carefully designed adaptive attacks, e.g., the attack success rate drops to 0% on CIFAR10.