Innovations in Next-Generation Sequencing are enabling generation of DNA sequence data at ever faster rates and at very low cost. Large sequencing centers typically employ hundreds of such systems. Such high-throughput and low-cost generation of data underscores the need for commensurate acceleration in downstream computational analysis of the sequencing data. A fundamental step in downstream analysis is mapping of the reads to a long reference DNA sequence, such as a reference human genome. Sequence mapping is a compute-intensive step that accounts for more than 30% of the overall time of the GATK workflow. BWA-MEM is one of the most widely used tools for sequence mapping and has tens of thousands of users. 
In this work, we focus on accelerating BWA-MEM through an efficient architecture aware implementation, while maintaining identical output. The volume of data requires distributed computing environment, usually deploying multicore processors. Since the application can be easily parallelized for distributed memory systems, we focus on performance improvements on a single socket multicore processor. BWA-MEM run time is dominated by three kernels, collectively responsible for more than 85% of the overall compute time. We improved the performance of these kernels by 1) improving cache reuse, 2) simplifying the algorithms, 3) replacing small fragmented memory allocations with a few large contiguous ones, 4) software prefetching, and 5) SIMD utilization wherever applicable - and massive reorganization of the source code enabling these improvements. 
As a result, we achieved nearly 2x, 183x, and 8x speedups on the three kernels, respectively, resulting in up to 3.5x and 2.4x speedups on end-to-end compute time over the original BWA-MEM on single thread and single socket of Intel Xeon Skylake processor. To the best of our knowledge, this is the highest reported speedup over BWA-MEM.

Efficient Architecture-Aware Acceleration of BWA-MEM for Multicore Systems

Natural graphs with skewed distributions raise unique challenges to distributed graph computation and partitioning. Existing graph-parallel systems usually use a “one-size-fits-all” design that uniformly processes all vertices, which either suffer from notable load imbalance and high contention for high-degree vertices (e.g., Pregel and GraphLab) or incur high communication cost and memory consumption even for low-degree vertices (e.g., PowerGraph and GraphX). In this article, we argue that skewed distributions in natural graphs also necessitate differentiated processing on high-degree and low-degree vertices. We then introduce PowerLyra, a new distributed graph processing system that embraces the best of both worlds of existing graph-parallel systems. Specifically, PowerLyra uses centralized computation for low-degree vertices to avoid frequent communications and distributes the computation for high-degree vertices to balance workloads. PowerLyra further provides an efficient hybrid graph partitioning algorithm (i.e., hybrid-cut) that combines edge-cut (for low-degree vertices) and vertex-cut (for high-degree vertices) with heuristics. To improve cache locality of inter-node graph accesses, PowerLyra further provides a locality-conscious data layout optimization. PowerLyra is implemented based on the latest GraphLab and can seamlessly support various graph algorithms running in both synchronous and asynchronous execution modes. A detailed evaluation on three clusters using various graph-analytics and MLDM (Machine Learning and Data Mining) applications shows that PowerLyra outperforms PowerGraph by up to 5.53X (from 1.24X) and 3.26X (from 1.49X) for real-world and synthetic graphs, respectively, and is much faster than other systems like GraphX and Giraph, yet with much less memory consumption. A porting of hybrid-cut to GraphX further confirms the efficiency and generality of PowerLyra.

PowerLyra: Differentiated Graph Computation and Partitioning on Skewed Graphs

Graphics Processing Units (GPUs) have numerous configuration and design options, including core frequency, number of parallel compute units (CUs), and available memory bandwidth. At many stages of the design process, it is important to estimate how application performance and power are impacted by these options. This paper describes a GPU performance and power estimation model that uses machine learning techniques on measurements from real GPU hardware. The model is trained on a collection of applications that are run at numerous different hardware configurations. From the measured performance and power data, the model learns how applications scale as the GPU's configuration is changed. Hardware performance counter values are then gathered when running a new application on a single GPU configuration. These dynamic counter values are fed into a neural network that predicts which scaling curve from the training data best represents this kernel. This scaling curve is then used to estimate the performance and power of the new application at different GPU configurations. Over an 8× range of the number of CUs, a 3.3× range of core frequencies, and a 2.9× range of memory bandwidth, our model's performance and power estimates are accurate to within 15% and 10% of real hardware, respectively. This is comparable to the accuracy of cycle-level simulators. However, after an initial training phase, our model runs as fast as, or faster than the program running natively on real hardware.

https://users.ece.utexas.edu/~derek/Papers/HPCA2015_GPUPowerModel.pdf

GPGPU performance and power estimation using machine learning

A data breach is the intentional or inadvertent exposure of confidential information to unauthorized parties. In the digital era, data has become one of the most critical components of an enterprise. Data leakage poses serious threats to organizations, including significant reputational damage and financial losses. As the volume of data is growing exponentially and data breaches are happening more frequently than ever before, detecting and preventing data loss has become one of the most pressing security concerns for enterprises. Despite a plethora of research efforts on safeguarding sensitive information from being leaked, it remains an active research problem. This review helps interested readers to learn about enterprise data leak threats, recent data leak incidents, various state‐of‐the‐art prevention and detection techniques, new challenges, and promising solutions and exciting opportunities. WIREs Data Mining Knowl Discov 2017, 7:e1211. doi: 10.1002/widm.1211

/pdf/enterprise-data-breach-causes-challenges-prevention-and-5fwnbecvya.pdf

Enterprise data breach: causes, challenges, prevention, and future directions

Mobile apps are notorious for collecting a wealth of private information from users. Despite significant effort from the research community in developing privacy leak detection tools based on data flow tracking inside the app or through network traffic analysis, it is still unclear whether apps and ad libraries can hide the fact that they are leaking private information. In fact, all existing analysis tools have limitations: data flow tracking suffers from imprecisions that cause false positives, as well as false negatives when the data flow from a source of private information to a network sink is interrupted; on the other hand, network traffic analysis cannot handle encryption or custom encoding. We propose a new approach to privacy leak detection that is not affected by such limitations, and it is also resilient to obfuscation techniques, such as encoding, formatting, encryption, or any other kind of transformation performed on private information before it is leaked. Our work is based on blackbox differential analysis, and it works in two steps: first, it establishes a baseline of the network behavior of an app; then, it modifies sources of private information, such as the device ID and location, and detects leaks by observing deviations in the resulting network traffic. The basic concept of black-box differential analysis is not novel, but, unfortunately, it is not practical enough to precisely analyze modern mobile apps. In fact, their network traffic contains many sources of non-determinism, such as random identifiers, timestamps, and server-assigned session identifiers, which, when not handled properly, cause too much noise to correlate output changes with input changes. The main contribution of this work is to make black-box differential analysis practical when applied to modern Android apps. In particular, we show that the network-based non-determinism can often be explained and eliminated, and it is thus possible to reliably use variations in the network traffic as a strong signal to detect privacy leaks. We implemented this approach in a tool, called Agrigento, and we evaluated it on more than one thousand Android apps. Our evaluation shows that our approach works well in practice and outperforms current state-of-the-art techniques. We conclude our study by discussing several case studies that show how popular apps and ad libraries currently exfiltrate data by using complex combinations of encoding and encryption mechanisms that other approaches fail to detect. Our results show that these apps and libraries seem to deliberately hide their data leaks from current approaches and clearly demonstrate the need for an obfuscation-resilient approach such as ours.

Obfuscation-resilient privacy leak detection for mobile apps through differential analysis

The leak of sensitive data on computer systems poses a serious threat to organizational security. Statistics show that the lack of proper encryption on files and communications due to human errors is one of the leading causes of data loss. Organizations need tools to identify the exposure of sensitive data by screening the content in storage and transmission, i.e., to detect sensitive information being stored or transmitted in the clear. However, detecting the exposure of sensitive information is challenging due to data transformation in the content. Transformations (such as insertion and deletion) result in highly unpredictable leak patterns. In this paper, we utilize sequence alignment techniques for detecting complex data-leak patterns. Our algorithm is designed for detecting long and inexact sensitive data patterns. This detection is paired with a comparable sampling algorithm, which allows one to compare the similarity of two separately sampled sequences. Our system achieves good detection accuracy in recognizing transformed leaks. We implement a parallelized version of our algorithms in graphics processing unit that achieves high analysis throughput. We demonstrate the high multithreading scalability of our data leak detection method required by a sizable organization.

/pdf/fast-detection-of-transformed-data-leaks-1zg00podb7.pdf

Fast Detection of Transformed Data Leaks

In the past, evaluating the architectural innovation of parallel computing devices relied on a benchmark suite based on existing programs, e.g., EEMBC or SPEC. However, with the growing ubiquity of parallel computing devices, we argue that it is unclear how best to express parallel computation, and hence, a need exists to identify a higher level of abstraction for reasoning about parallel application requirements. Therefore, the goal of this combination "Work-in-Progress and Vision" paper is to delineate application requirements in a manner that is not overly specific to individual applications or the optimizations used for certain hardware platforms, so that we can draw broader conclusions about hardware requirements. Our initial effort, dubbed "OpenCL and the 13 Dwarfs" or OCD for short, realizes Berkeley's 13 computational dwarfs of scientific computing in OpenCL, where each dwarf captures a pattern of computation and communication that is common to a class of important applications.

/pdf/opencl-and-the-13-dwarfs-a-work-in-progress-1tt6icqya8.pdf

OpenCL and the 13 dwarfs: a work in progress

Computational biology sequence alignment tools using the Burrows-Wheeler Transform (BWT) are widely used in next-generation sequencing (NGS) analysis. However, despite extensive optimization efforts, the performance of these tools still cannot keep up with the explosive growth of sequencing data. Through an in-depth performance analysis of BWA, a popular BWT-based aligner on multicore architectures, we demonstrate that such tools are limited by memory bandwidth due to their irregular memory access patterns. We then propose a locality-aware implementation of BWA that aims at optimizing its performance by better exploiting the caching mechanisms of modern multicore processors. Experimental results show that our improved BWA implementation can reduce last-level cache (LLC) misses by 30% and translation look aside buffer (TLB) misses by 20%, resulting in up to 2.6-fold speedup over the original BWA implementation.

/pdf/optimizing-burrows-wheeler-transform-based-sequence-56q3j28j4j.pdf

Optimizing burrows-wheeler transform-based sequence alignment on multicore architectures

BLAST, short for Basic Local Alignment Search Tool, is a ubiquitous tool used in the life sciences for pairwise sequence search. However, with the advent of next-generation sequencing (NGS), whether at the outset or downstream from NGS, the exponential growth of sequence databases is outstripping our ability to analyze the data. While recent studies have utilized the graphics processing unit (GPU) to speedup the BLAST algorithm for searching protein sequences (i.e., BLASTP), these studies use coarse-grained parallelism, where one sequence alignment is mapped to only one thread. Such an approach does not efficiently utilize the capabilities of a GPU, particularly due to the irregularity of BLASTP in both execution paths and memory-access patterns. To address the above shortcomings, we present a fine-grained approach to parallelize BLASTP, where each individual phase of sequence search is mapped to many threads on a GPU. This approach, which we refer to as cuBLASTP, reorders data-access patterns and reduces divergent branches of the most time-consuming phases (i.e., hit detection and ungapped extension). In addition, cuBLASTP optimizes the remaining phases (i.e., gapped extension and alignment with trace back) on a multicore CPU and overlaps their execution with the phases running on the GPU.

/pdf/cublastp-fine-grained-parallelization-of-protein-sequence-4ctn5bavi9.pdf

cuBLASTP: Fine-Grained Parallelization of Protein Sequence Search on CPU+GPU

BLAST, short for Basic Local Alignment Search Tool, is a fundamental algorithm in the life sciences that compares biological sequences. However, with the advent of next-generation sequencing (NGS) and increase in sequence read-lengths, whether at the outset or downstream from NGS, the exponential growth of sequence databases is arguably outstripping our ability to analyze the data. Though several recent studies have utilized the graphics processing unit (GPU) to speedup the BLAST algorithm for searching protein sequences (i.e., BLASTP), these studies used coarse-grained parallel approaches, where one sequence alignment is mapped to only one thread. Moreover, due to the irregular memory access patterns in BLASTP, there remain significant challenges to map the most time-consuming phases (i.e., hit detection and ungapped extension) to the GPU using a fine-grained multithreaded approach. To address the above issues, we propose cuBLASTP, an efficient fine-grained BLASTP implementation for the GPU using CUDA. Our cuBLASTP realization encompasses many research contributions, including (1) memory-access reordering to reorder hits from column-major order to diagonal-major order, (2) position-based indexing to map a hit with a packed data structure to a bin, (3) aggressive hit filtering to eliminate hits beyond the threshold distance along the diagonal, (4) diagonal-based parallelism and hit-based parallelism for ungapped extension to extend sequences with different lengths in databases, and (5) hierarchical buffering to reduce memory-access overhead for the core data structures. The experimental results show that on a NVIDIA Kepler GPU, cuBLASTP delivers up to a 5.0-fold speedup over sequential FSA-BLAST and a 3.7-fold speedup over multithreaded NCBI-BLAST for the overall program execution. In addition, compared with GPU-BLASTP (the fastest GPU implementation of BLASTP to date), cuBLASTP achieves up to a 2.8-fold speedup for the kernel execution on the GPU and a 1.8-fold speedup for the overall program execution.

/pdf/cublastp-fine-grained-parallelization-of-protein-sequence-41upvksw76.pdf

Jing Zhang

Papers

Fast Detection of Transformed Data Leaks

OpenCL and the 13 dwarfs: a work in progress

Optimizing burrows-wheeler transform-based sequence alignment on multicore architectures

cuBLASTP: Fine-Grained Parallelization of Protein Sequence Search on CPU+GPU

cuBLASTP: Fine-Grained Parallelization of Protein Sequence Search on a GPU