high performance computing on graphics processing units: hgpu.org

Ensuring high productivity in scientific software development necessitates developing and maintaining a single codebase that can run efficiently on a range of accelerator-based supercomputing platforms. While prior work has investigated the performance portability of a few selected proxy applications or programming models, this paper provides a comprehensive study of a range of proxy applications implemented […]

Since the advent of parallel algorithms in the C++17 Standard Template Library (STL), the STL has become a viable framework for creating performance-portable applications. Given multiple existing implementations of the parallel algorithms, a systematic, quantitative performance comparison is essential for choosing the appropriate implementation for a particular hardware configuration. In this work, we introduce a […]

Vision Transformer (ViT) models which were recently introduced by the transformer architecture have shown to be very competitive and often become a popular alternative to Convolutional Neural Networks (CNNs). However, the high computational requirements of these models limit their practical applicability especially on low-power devices. Current state-of-the-art employs approximate multipliers to address the highly increased […]

We introduce QUICK, a group of novel optimized CUDA kernels for the efficient inference of quantized Large Language Models (LLMs). QUICK addresses the shared memory bank-conflict problem of state-of-the-art mixed precision matrix multiplication kernels. Our method interleaves the quantized weight matrices of LLMs offline to skip the shared memory write-back after the dequantization. We demonstrate […]

CodeCompose is an AI-assisted code authoring tool powered by large language models (LLMs) that provides inline suggestions to 10’s of thousands of developers at Meta. In this paper, we present how we scaled the product from displaying single-line suggestions to multi-line suggestions. This evolution required us to overcome several unique challenges in improving the usability […]

The rapid development of large language models has revolutionized code intelligence in software development. However, the predominance of closed-source models has restricted extensive research and development. To address this, we introduce the DeepSeek-Coder series, a range of open-source code models with sizes from 1.3B to 33B, trained from scratch on 2 trillion tokens. These models […]

As of 2023 we have taught more than 24,000 students over the course of 106 events using the Wide Area Classroom, a novel distributed teaching platform. This has been a successful effort gauged by several important metrics. We describe both the technical and logistical structure of these events as well as the specific HPC curriculums […]

Adjusting batch sizes and adaptively tuning other hyperparameters can significantly speed up deep neural network (DNN) training. Despite the ubiquity of heterogeneous clusters, existing adaptive DNN training techniques solely consider homogeneous environments. Optimizing distributed DNN training over heterogeneous clusters is technically challenging, and directly adapting existing techniques results in low utilization and poor performance. To […]

Virtual screening is an early stage in the drug discovery process that selects the most promising candidates. In the urgent computing scenario, finding a solution in the shortest time frame is critical. Any improvement in the performance of a virtual screening application translates into an increase in the number of candidates evaluated, thereby raising the […]

Dynamic memory management is critical for efficiently porting modern data processing pipelines to GPUs. However, building a general-purpose dynamic memory manager on GPUs is challenging due to the massive parallelism and weak memory coherence. Existing state-of-the-art GPU memory managers, Ouroboros and Reg-Eff, employ traditional data structures such as arrays and linked lists to manage memory […]

A heterogeneous computing system is a system composed of different types of computing units. SYCL is a software development framework with which programs can be developed for such systems. It uses the concept of kernels, where a kernel executes code inside it in parallel, and different kernels can be executed concurrently on multiple computing units. […]

This paper describes LeftoverLocals: a vulnerability that allows data recovery from GPU memory created by another process on Apple, Qualcomm, and AMD GPUs. LeftoverLocals impacts the security posture of GPU applications, with particular significance to LLMs and ML models that run on impacted GPUs. By recovering local memory, an optimized GPU memory region, we built […]