AMAX, a leading provider of turnkey rack-scale High Performance Computing (HPC) solutions, Deep Learning/AI applications and server appliance manufacturing, announces the new AceleMax X-122-Flex server solution featuring Intel’s next-generation Data Center GPU Flex Series, (formerly code-named Arctic Sound-M), providing the capability of a graphics processing (GPU) solution handling high density and complex workloads targeted towards media delivery, cloud gaming, AI, metaverse, and other emerging visual cloud use cases.
AMAX provides the flexibility to choose the right fit for specific workloads through offering the new Data Center GPU Flex Series in two variants. The 150W variant includes 32 Xe cores in a three-quarter length full-height PCIe Gen4 card and is ideal for more complex AI models, including multiple object detection and multiple classification models. The 75W option comes as a low-profile PCIe Gen4 card with a total of 16 Xe cores and is suitable for media delivery and lighter AI models, such as simple object detection. Both configurations include four Xe media engines, GDDR6 memory, ray tracing units, built-in XMX AI acceleration, and boast the industry’s first AV1 hardware encoder.
The AceleMax X-112-Flex, engineered by AMAX, is a 1U server that features dual Intel 3rd Gen Xeon Processors with up to 80 cores, two Intel Data Center Flex Series GPUs, up to 12TB of DDR4 memory, and 12 Hot-Swap 2.5x SSD front drives.
Intel’s Data Center GPU Flex Series is powered by Intel’s Xe-HPG architecture, which is designed for AI inference, Android and Windows cloud gaming, video transcoding, and virtual desktop infrastructure (VDI) applications. The Data Center Flex Series is a versatile GPU with industry-leading transcode quality and performance, with a target performance of 150 trillion operations per second (TOPS). A single solution designed to handle a diverse range of workloads without sacrificing performance or quality, while minimizing total cost of ownership (TCO).
Another advantage of the Data Center GPU Flex Series is that it is supported by a full solution software stack that provides developers with an open-source library for streaming media, cloud gaming, and inference, as well as APIs, frameworks, and the most recent codecs. By implementing oneAPI, an open alternative to proprietary language lock-in that expands full hardware performance, developers are enabled to design open, portable code that will take full advantage of any CPU and GPU combination.
Intel’s Data Center GPU Flex Series cards will not be available through standard component distributors and will only be available through solutions such as the AMAX AceleMax X-112-Flex. AMAX’s AceleMax family of advanced GPU platforms continues to leverage the latest industry accelerator technologies to deliver optimized data center performance from the node to the rack.
Sign up for the free insideBIGDATA newsletter.
Join us on Twitter: @InsideBigData1 – https://twitter.com/InsideBigData1
*
*
document.getElementById( “ak_js_1” ).setAttribute( “value”, ( new Date() ).getTime() );
In this special guest feature, Peter Bernard, CEO, Datagration, discusses how oil and gas industry predictive maintenance doesn’t just provide an economic value, it also boosts safety by anticipating unpredicted failures amongst aging infrastructure in many rigs. By applying machine learning (ML) to pumps and compressors, operators can easily detect equipment failure before it happens, greatly reducing downtime and extending the lifespan of machinery. Maintenance strategies across upstream oil and gas require a smarter approach, and ML provides the answer.
A new guidebook from Dataiku explores the basic concepts of machine learning. Learn about the common algorithms used in machine learning today as well as how to evaluate and fine-tune these models. Download the full report to discover why machine learning is growing and what it can potentially do for your business. 
Copyright © 2022