nvCOMP: a CUDA library for Fast Lossless Compression and Decompression on GPUs

Hello CUDA community,

We're happy to announce our next online meetup!
On October 5th (8 am (PDT), 18:00 (MSK - UTC+3)), we'll talk about nvCOMP: a CUDA library for Fast Lossless Compression and Decompression on the GPU. We will add a link to the Zoom meeting closer to the event so stay tuned.

Speaker:
Nikolay Sakharnykh

Bio:
Nikolay Sakharnykh is a senior AI developer technology manager at NVIDIA. He leads a team working on accelerating data analytics and machine learning applications on GPUs. He is interested in novel memory management techniques.

Abstract:
Modern data analytics and machine learning applications manage large amounts of data. Often the interconnect between the processors and storage systems is the main bottleneck. Lossless data compression is one of the most common approaches to alleviate data transfer bottleneck. However, traditional compression methods were designed decades ago and are difficult to parallelize on massively-parallel architectures like GPUs. In this talk I'll introduce nvCOMP - a new CUDA library for lossless data compression and decompression that's optimized for the GPU. nvCOMP supports standard compression schemes like LZ4 and Snappy, as well as some novel methods designed for better parallelism. The library provides a light-weight C/C++ interface across all the compressors for easier integration into applications and compatibility with standard CPU libraries like liblz4. I'll cover performance results, key ideas behind the GPU implementations and a few exciting use cases. Check out the library before the talk and come prepared with questions! GitHub: https://github.com/NVIDIA/nvcomp , NVIDIA developer page: https://developer.nvidia.com/nvcomp

Agenda:

• Welcome message
• Raffle of free CLion licenses
• The talk
• Post meetup discussions