- IBM's quickest mainframe laptop up to now has a Telum II CPU with on-chip AI coprocessor
- There’s additionally an AI accelerator that it hopes will make it a gorgeous choice
- The IBM z17 brings Large Blue within the age of AI however will this be sufficient to fend off competitors?
IBM has introduced the z17, a brand new mainframe to deal with rising AI calls for on enterprise infrastructure.
Positioned as a basis for hybrid cloud environments, and with help for real-time AI and enterprise-grade resilience, the IBM z17 is designed to deal with transaction-heavy workloads, enhance operational effectivity, and tackle safety considerations in industries with stringent compliance wants.
Central to the brand new mainframe is the Telum II processor, which was initially introduced at Sizzling Chips 2024. Developed utilizing Samsung 5nm expertise, it integrates an on-chip AI coprocessor to help inferencing duties, together with small language fashions with fewer than 8 billion parameters.
Large on safety
Alongside the processor, Large Blue plans to supply the Spyre Accelerator card (additionally initially previewed at Sizzling Chips) to enhance the Telum II and lengthen the AI compute capabilities for unstructured knowledge processing equivalent to text-based generative AI.
The z17 can accommodate as much as 48 of those accelerator playing cards, permitting scalability throughout enterprise workloads. It’s anticipated to be obtainable 4Q 2025 by way of PCIe card.
Safety is an enormous focus for the IBM z17 and consists of AI-powered options equivalent to Delicate Knowledge Tagging for z/OS and IBM Risk Detection for z/OS, each of which use pure language processing to establish and shield delicate knowledge or scan for potential threats.
As well as, it helps NIST-standardized quantum-safe cryptographic algorithms to deal with future regulatory necessities.
Signal as much as the TechRadar Professional publication to get all the highest information, opinion, options and steering your small business must succeed!
The z17 system additionally incorporates a brand new knowledge processing unit to speed up I/O protocols for networking and storage.
IBM says it expects software builders to profit from AI-driven assistants that may automate duties throughout the software program growth lifecycle, bettering productiveness and decreasing talent transition points in mainframe environments.
Transactional AI use circumstances equivalent to fraud detection, cash laundering prevention, and anomaly detection can now be deployed nearer to the info supply, IBM says, with help for multimodel inference to enhance accuracy and cut back false positives.
"The business is rapidly studying that AI will solely be as priceless because the infrastructure it runs on," stated Ross Mauri, basic supervisor of IBM Z and LinuxONE, IBM.
"With z17, we're bringing AI to the core of the enterprise with the software program, processing energy, and storage to make AI operational rapidly. Moreover, organizations can put their huge, untapped shops of enterprise knowledge to work with AI in a secured, cost-effective means."
Accessible in configurations that help as much as 208 processors and 64TB of reminiscence, the z17, which is the fruits of 5 years of design and growth, is designed to function at 5.5GHz and comes housed in as much as 4 frames. Whereas aimed toward vital workloads, IBM can be positioning it as half of a bigger hybrid cloud technique.
IBM additionally took the wraps off z/OS 3.2, the following model of its flagship working system for IBM Z methods. That is deliberate for launch within the third quarter of 2025.
You may additionally like
- Open supply tech permits customers to 'timeshare' GPU assets for AI functions
- How IBM is shaping the way forward for quantum computing
- 'Information and group' – IBM and the advantages of the AI Alliance one 12 months on