NeuroBlade Empowers Accelerated Next-Gen Data Analytics

NeuroBlade, Inc. announces its integration with the newly released Amazon Elastic Compute Cloud (Amazon EC2) F2 instances, powered by AMD FPGA and EPYC CPU technologies. This collaboration brings performance and efficiency to cloud-native data analytics workloads, leveraging NeuroBlade’s analytics acceleration technology.

Deploying NeuroBlade Acceleration technology on EC2 F2 instances provides a cutting-edge cloud-based solution that aligns with the growing trend of data migration to the cloud. With most organizations managing their data in cloud environments, offering access to data analytics-optimized instances within this infrastructure is essential-especially in the era of AI and ML, where efficiently processing vast amounts of data is critical.

NeuroBlade’s data analytics acceleration technology is available on the AWS cloud via the newly released EC2 F2 instances or on-premises as a standard PCIe card that can be installed into existing server infrastructures. This flexibility empowers customers to integrate NeuroBlade’s cutting-edge acceleration into their data platforms, enabling them to test and deploy this advanced technology close to their data and unlock faster decision-making, improved outcomes, and enhanced scalability.

“The integration of NeuroBlade’s analytics acceleration technology with Amazon EC2 F2 instances represents a pivotal shift in data processing. By leveraging AMD FPGA and EPYC CPU capabilities, we’re enabling businesses to overcome the limitations of traditional CPU-centric systems, delivering faster queries, improved throughput, and reduced TCO. This collaboration transforms how organizations handle cloud-native analytics, unlocking unprecedented performance and scalability for data-intensive workloads.” said Elad Sity, CEO and co-founder, NeuroBlade.

The availability of NeuroBlade’s Acceleration technology on EC2 F2 instances enables customers to quickly evaluate its benefits, including faster query processing, TCO savings, and the ability to enhance their services by utilizing more efficient data analytics processing. The company’s solution integrates with popular open-source query engines like Presto and Apache Spark, delivering market-leading query throughput efficiency (QpH/$). Customers can now unlock scalability and performance for their data analytics workloads.

NeuroBlade also provides reference integrations of its accelerator technology with Prestissimo (Presto + Velox) and Apache Spark. These setups enable customers to run industry-standard benchmarks such as TPC-H and TPC-DS or test their workloads with their own datasets. This facilitates apples-to-apples comparisons between accelerated and non-accelerated implementations on the same cloud infrastructure, clearly demonstrating the performance gains and cost efficiencies of NeuroBlade’s Acceleration technology in comparison to native vectorized processing on the CPU.

Comments

The company is very confidential, almost invisible, probably due to marketing, positioning and go-to-market challenges even with a change at the top of marketing a few quarters ago. It appears to be an engineering driven entity. The company was present at FMS 2023 on the Kioxia booth but was absent of the show this year.

Some executives are no longer listed on their website such as Lior Genzel, previous chief business officer, or Krishna Maheshwari, chief product officer.

So the question of the need of such solutions is on the table and the complexity it may create especially if users would consider multiple xPU for different usages in same server chassis.

We saw so many stories that hit the wall or finish in the basket for a small amount. And at the same time for some needs, it makes sense especially for network acceleration with the SmartNIC approach and 2 good examples are Annapurna Labs picked by Amazon in 2015 for $350 million or Mellanox bought by Nvidia in 2019 for $6.9 billion, the last one is of course not only for the BlueField xPU. Where is Nebulon today? Fungible has been acquired by Microsoft for less the total money they raised. Pensando was targeted by AMD for a big amount, $1.9 billion, in 2022. Big moves also are present like AMD again who acquired Xilinx for $35 billion in 2021 but it goes beyond xPU but converges towards the same notion as it covers off-load data processing. Swarm64 has been swallowed by ServiceNow in 2021.

So we'll see where NeuroBlade's trajectory will go, we have some doubts...