1,184 papers · page 16 of 60
In this paper, we propose a runtime, called HUM, which hides host-to-device memory copy time without any code modification. It overlaps the host-to-device memory copy with host computation or CUDA kernel computation by exploiting Unified Memory and fault mechanisms. HUM provides …
Multi-producer multi-consumer FIFO queue is one of the fundamental concurrent data structures used in software systems. A lot of progress has been done on designing concurrent bounded and unbounded queues [1--10]. As previous works show, it is extremely hard to come up with an ef…
Concurrent programming can be notoriously complex and error-prone. Programming bugs can arise from a variety of sources, such as operation re-reordering, or incomplete understanding of the memory model. A variety of formal and model checking methods have been developed to address…
Load imbalance pervasively exists in distributed deep learning training systems, either caused by the inherent imbalance in learned tasks or by the system itself. Traditional synchronous Stochastic Gradient Descent (SGD) achieves good accuracy for a wide variety of tasks, but rel…
Tensor computations present significant performance challenges that impact a wide spectrum of applications. Efforts on improving the performance of tensor computations include exploring data layout, execution scheduling, and parallelism in common tensor kernels. This work present…
Data movement has a significant impact on program performance. For multithread programs, this impact is amplified, since different threads often interfere with each other by competing for shared cache space. However, recent de facto locality metrics consider either sequential exe…
Hierarchical matrix approximations have gained significant traction in the machine learning and scientific community as they exploit available low-rank structures in kernel methods to compress the kernel matrix. The resulting compressed matrix, HMatrix, is used to reduce the comp…
In this paper, we present the first comprehensive performance characterization and optimization of ARM barriers on both mobile and server platforms. We draw a set of observations through several abstracted models and validate them in scenarios where barriers are intensively used.…
Efficient ordered in-memory key-value (KV-)maps are paramount for the scalability of modern data platforms. In managed languages like Java, KV-maps face unique challenges due to the high overhead of garbage collection (GC). We present Oak, a scalable concurrent KV-map for enviro…
In this paper, we present a universal memory reclamation scheme, Wait-Free Eras (WFE), for deleted memory blocks in wait-free concurrent data structures. WFE's key innovation is that it is completely wait-free. Although some prior techniques provide similar guarantees for certain…
Queues are fundamental concurrent data structures, but despite years of research, even the state-of-the-art queues scale poorly. This poor scalability occurs because of contended atomic read-modify-write (RMW) operations. This paper makes a first step towards designing a scalabl…
Sparse general matrix-matrix multiplication on GPUs is challenging due to the varying sparsity patterns of sparse matrices. Existing solutions achieve good performance for certain types of matrices, but fail to accelerate all kinds of matrices in the same manner. Our approach com…
Data sketches are approximate succinct summaries of long data streams. They are widely used for processing massive amounts of data and answering statistical queries about it. Existing libraries producing sketches are very fast, but do not allow parallelism for creating sketches u…
May-Happen-in-Parallel (MHP) analysis forms the basis for many problems of program analysis and program understanding. MHP analysis can also be used by IDEs (integrated-development-environments) to help programmers to refactor parallel-programs, identify racy programs, understand…
We show how to exploit graph sparsity in the Floyd-Warshall algorithm for the all-pairs shortest path (Apsp) problem. Floyd-Warshall is an attractive choice for Apsp on high-performing systems due to its structural similarity to solving dense linear systems and matrix multiplicat…
Many big data processing applications rely on a top-k retrieval building block, which selects (or approximates) the k highest-scoring data items based on an aggregation of features. In web search, for instance, a document's score is the sum of its scores for all query terms. Top-…
Hardware and software faults increasingly surface in today's computing environment and vast theoretical and practical research efforts are being devoted to overcoming the effects of malfunctionality in the computing process. Most research to date, however, has focused on how to d…
As the widening gap between GPU computing capability and other components (CPU, PCIe bus and communication network), it's increasingly challenging to design high performance parallel algorithms for large CPU-GPU heterogeneous systems. There are mainly two reasons. Firstly, simply…
While there has been significant work on parallel graph processing, there has been very surprisingly little work on high-performance hypergraph processing. This paper presents a collection of efficient parallel algorithms for hypergraph processing, including algorithms for betwee…
We present XIndex, a concurrent ordered index designed for fast queries. Similar to a recent proposal of the learned index, XIndex uses learned models to optimize index efficiency. Comparing with the learned index, XIndex is able to effectively handle concurrent writes without af…