Sign up for our newsletter and get the latest HPC news and analysis.
Send me information from insideHPC:


Mellanox to Accelerate EDA Workloads with WekaIO

Today WekaIO announced that Mellanox has selected the WekaIO Matrix file system to accelerate application performance scaling for their EDA workloads.

Mellanox is a forward-thinking company, we continuously improve our workflows by incorporating new deep learning techniques into our data-intensive EDA workloads, so we needed a modern file system to support our initiatives,” said Udi Weinstein, Chief Information Officer of Mellanox. “The integration of WekaIO’s Matrix software in our infrastructure increases the efficiency of our datacenter, keeping pace with our application performance requirements while delivering exascale capacity at the best economics. With our current IT infrastructure, we are well-positioned to deliver the best IC products to the market in to the future.”

Mellanox interconnect solutions increase datacenter efficiency by providing the highest throughput and lowest latency, delivering data faster to applications, unlocking system performance and improving data security. Mellanox offers a choice of fast interconnect products: adapters, switches, software and silicon that accelerate application performance and maximize business results for a wide range of markets including cloud and hyperscale, HPC, AI, enterprise data centers, cyber security, storage, financial services and more.

With a single namespace that spans local storage and the cloud, Matrix delivers performance that is 3x that of local file systems and 10x that of traditional NAS. The Matrix software is a highly resilient parallel file system that delivers the highest bandwidth, lowest latency performance to any InfiniBand or Ethernet enabled GPU or CPU based cluster. The POSIX-compliant file system leverages an innovative, highly optimized software stack that leverages modern architectures such as low-latency NVMe-oF and massively distributed metadata. Built for enterprise grade HPC, it tightly integrates object storage for best economics at scale and supports features such as snapshots, cloud backup and multiprotocol support including NFS and SMB.

In addition to Mellanox using Matrix for its own EDA workflow, Mellanox and WekaIO have many joint customer deployments with Deep Learning (DL) workflows. EDA benefits from DL on many levels, be it product innovation, chip design or lower costs, and yet DL can be computationally intensive, especially the inference phase, and as such Matrix has become the go-to file system of choice for enterprises’ supporting DL workloads. Mellanox ConnectX 100G network, and soon 200G, was also a key component of the Matrix cluster design that broke performance records on both the SPEC SFS and IO500 10 Node Challenge benchmarks.

Cultivating a full-spectrum business relationship with companies like Mellanox benefits the entire HPC ecosystem of use cases, not just EDA, but also AI, Machine Learning, Finance, Life Sciences and Genomics, and M&E,” said Liran Zvibel, co-founder and CEO of WekaIO. “We have many synergies with Mellanox, from the markets we target to our core value proposition of accelerating application performance. Not only are we proving we deliver the best performance and value for EDA in the Mellanox datacenter, but we jointly enable many deep learning workloads in successful customer deployments as well. And together we have successfully proven we deliver the best performance on the HPC and Machine Learning industries leading benchmarks. The Mellanox leading interconnect solutions and Matrix file system are an undeniably, unbeatable storage solution for data intensive HPC.”

Sign up for our insideHPC Newsletter

Leave a Comment

*

Resource Links: