chevron_leftBack

US - English China - 简体中文 India - English Japan - 日本語 Malaysia - English Singapore - English Taiwan – 繁體中文
Micron Insight Micron Blog
Storage // Innovations

Walt Hubis on NVM Express® 2.0

By Rebecca Lewington - 2021-07-06
Categories Communities Company Events Executive Insight Corner Innovations Memory Storage
Products Advanced Computing Solutions Advanced Solutions DRAM DRAM Modules High Bandwidth Memory Managed NAND Memory Memory Cards Multichip Packages NAND Flash NOR Flash SpecTek SSDs Storage Storage Platforms Ultra-Bandwidth Solutions

NVM Express, or NVMe®, defines how computers talk to SSDs. It’s a fundamental part of modern computing systems, from laptops to data centers to supercomputers. So the recent release of version 2.0 must be a big deal. But why?

Micron is a strong supporter of the NVMe 2.0 specification and has been a contributor to the development of many of its new features. These features open up really interesting possibilities for application developers.

Walt Hubis represents Micron on the committee that determines what makes it into each version of NVMe, so he’s the ideal person to help me dive into the new features of NVMe 2.0 and the standards process itself.

For a deeper dive on what’s new in NVMe 2.0, read Walt’s blog post.

Learn more at micron.com/nvmes.

Rebecca Lewington

Rebecca Lewington

Technology Evangelist, Global Corporate Marketing and Corporate Communications
Rebecca is responsible for telling Micron’s technology story to show how fractally innovative our people are and how we’re enabling the world to harness data to accelerate exploration and discovery. Based in California, she holds a master’s degree in engineering from the University of London and has been granted 15 patents in the field of semiconductor plasma etch manufacturing systems.

Continue reading

Related posts

Storage

The Micron 9400 NVMe SSD is the top PCIe Gen4 SSD for AI Storage

2023-09-19
MLCommons recently published the first round of MLPerf benchmark submissions – the first storage benchmark focused on AI Training workloads. The Micron 9400 NVMe SSD achieved the top PCIe Gen4 SSD score, supporting 17 emulated AI accelerators and drives 41 samples per second training throughput of a Three Dimensional Medical Imaging model (Unet3D).
Read more >
Storage

Identifying latency outliers in workload testing

2023-08-17
Sometimes there are large latency spikes when running and collecting workload traces of RocksDB. To emulate this issue, we used Flexible IO tester (FIO), a relatively simple tool to use that enables the user to see the latency spikes. In this blog, we discuss the methods used to identify the root cause of latency spikes in a mixed read and write workload.
Read more >
Innovations // Memory // Storage

Storage for AI Training: MLPerf Storage on the Micron® 9400 NVMe™ SSD

2023-08-17
What are the common storage requirements for AI workloads? MLPerf Storage aims to develop a process for testing exactly that. Their approach utilizes the Deep Learning I/O (DLIO) Benchmark which uses similar data loaders as real workloads (pytorch, tensorflow, etc.) to move data from storage to CPU memory. In this blog, we use DLIO to examine the IO pattern of the Unet 3D AI training model on the Micron 9400 NVMe SSD.
Read more >
+