Nikolay Sakharnykh

Nikolay Sakharnykh is a senior AI developer technology manager at NVIDIA. He leads an engineering team partnering with developers across the world to bring the best possible performance for their data analytics and machine learning applications on GPU accelerated computing systems. He has over 14 years of experience in developing parallel algorithms and performance optimizations for graphics, HPC, and AI.
Avatar photo

Posts by Nikolay Sakharnykh

Simulation / Modeling / Design Aug 22, 2023

Simplifying GPU Application Development with Heterogeneous Memory Management

Heterogeneous Memory Management (HMM) is a CUDA memory management feature that extends the simplicity and productivity of the CUDA Unified Memory programming... 16 MIN READ
Data Science Mar 06, 2023

Maximizing Performance with Massively Parallel Hash Maps on GPUs

Decades of computer science history have been devoted to devising solutions for efficient storage and retrieval of information. Hash maps (or hash tables) are a... 19 MIN READ
Simulation / Modeling / Design Mar 18, 2022

Accelerating Lossless GPU Compression with New Flexible Interfaces in NVIDIA nvCOMP

Compression can improve performance in a variety of use cases such as DL workloads, databases, and general HPC. On the GPU, compression can accelerate inter-GPU... 12 MIN READ
Data Center / Cloud Oct 05, 2021

Improving GPU Memory Oversubscription Performance

Since its introduction more than 7 years ago, the CUDA Unified Memory programming model has kept gaining popularity among developers. Unified Memory provides a... 16 MIN READ
Conversational AI / NLP Dec 18, 2020

Optimizing Data Transfer Using Lossless Compression with NVIDIA nvcomp

One of the most interesting applications of compression is optimizing communications in GPU applications. GPUs are getting faster every year. For some apps,... 17 MIN READ
Simulation / Modeling / Design Apr 15, 2020

Introducing Low-Level GPU Virtual Memory Management

There is a growing need among CUDA applications to manage memory as quickly and as efficiently as possible. Before CUDA 10.2, the number of options available to... 23 MIN READ