Lecture

Hardware Interconnect Optimization: RDMA and GPU Integration

Graph Chat

Chat with Graph Search

Ask any question about EPFL courses, lectures, exercises, research, news, etc. or try the example questions below.

DISCLAIMER: The Graph Chatbot is not programmed to provide explicit or categorical answers to your questions. Rather, it transforms your questions into API requests that are distributed across the various IT services officially administered by EPFL. Its purpose is solely to collect and recommend relevant references to content that you can explore to help you answer your questions.

Description

This lecture discusses the optimization of hardware interconnects, focusing on RDMA and GPU integration. Topics include memory scalability, storage disaggregation, heterogeneous device interconnection, RDMA in GPU-enabled servers, and execution traits in heterogeneous servers. It also covers observability, load balancing, offloading tasks to remote nodes, reducing data transfer trips, and representing operations in plans. The experimental setup with CPU-only and GPU-only servers is explored, along with the performance comparison in SSB benchmarks. The lecture concludes with strategies for moving complexity to the GPU side of the interconnect and utilizing CPU as a data cache.

This video is available exclusively on Mediaspace for a restricted audience. Please log in to MediaSpace to access it if you have the necessary permissions.

Watch on Mediaspace

Official source

Related concepts (21)

Heterogeneous System Architecture

Heterogeneous System Architecture (HSA) is a cross-vendor set of specifications that allow for the integration of central processing units and graphics processors on the same bus, with shared memory and tasks. The HSA is being developed by the HSA Foundation, which includes (among many others) AMD and ARM. The platform's stated aim is to reduce communication latency between CPUs, GPUs and other compute devices, and make these various devices more compatible from a programmer's perspective, relieving the programmer of the task of planning the moving of data between devices' disjoint memories (as must currently be done with OpenCL or CUDA).

Heterogeneous computing

Heterogeneous computing refers to systems that use more than one kind of processor or core. These systems gain performance or energy efficiency not just by adding the same type of processors, but by adding dissimilar coprocessors, usually incorporating specialized processing capabilities to handle particular tasks. Usually heterogeneity in the context of computing referred to different instruction-set architectures (ISA), where the main processor has one and other processors have another - usually a very different - architecture (maybe more than one), not just a different microarchitecture (floating point number processing is a special case of this - not usually referred to as heterogeneous).

Graphics processing unit

A graphics processing unit (GPU) is a specialized electronic circuit initially designed to accelerate computer graphics and (either on a video card or embedded on the motherboards, mobile phones, personal computers, workstations, and game consoles). After their initial design, GPUs were found to be useful for non-graphic calculations involving embarrassingly parallel problems due to their parallel structure. Other non-graphical uses include the training of neural networks and cryptocurrency mining.

Larrabee (microarchitecture)

Larrabee is the codename for a cancelled GPGPU chip that Intel was developing separately from its current line of integrated graphics accelerators. It is named after either Mount Larrabee or Larrabee State Park in Whatcom County, Washington, near the town of Bellingham. The chip was to be released in 2010 as the core of a consumer 3D graphics card, but these plans were cancelled due to delays and disappointing early performance figures.

Coprocessor

A coprocessor is a computer processor used to supplement the functions of the primary processor (the CPU). Operations performed by the coprocessor may be floating-point arithmetic, graphics, signal processing, string processing, cryptography or I/O interfacing with peripheral devices. By offloading processor-intensive tasks from the main processor, coprocessors can accelerate system performance. Coprocessors allow a line of computers to be customized, so that customers who do not need the extra performance do not need to pay for it.

Related lectures (26)

Accelerator-centric OS Architecture

By the instructor Mark Silberstein explores accelerator-centric OS architecture and the role of CPUs in modern computers.

GPU Memory Hierarchy and Optimization

Explores GPU memory hierarchy, optimization challenges, and algorithm efficiency.

GPU Advanced: Memory Management and Optimization

Explores advanced GPU memory management, CUDA programming, and data-parallel computing for optimizing performance.

Fast Interconnects: Scalable Co-processing with GPUs Log in to Mediaspace to watch this video

Explores the use of fast interconnects for scalable co-processing with GPUs in databases, emphasizing the importance of overcoming the transfer bottleneck and reevaluating assumptions for performance improvements.

GPUs: Multithreading and Architecture