Integrated AI accelerators for near real-time analytics and enhanced performance on the IBM Z® platform
Driving next-gen performance and AI
The IBM® Telum® processor is an advanced processor that powers many of the features and benefits of the IBM z16®. Its successor, the Telum II powering the new IBM z17®, builds on this foundation with enhanced performance, increased memory capacity and improved AI acceleration. It is designed to help customers achieve business insights at scale across banking, finance, trading, insurance applications and customer interactions.
Telum features on-chip AI acceleration for real-time analytics and decision-making without extra hardware. Telum II boosts AI performance with 4x the compute power and connected processing clusters.3
The Telum processor supports advanced encryption, including quantum-safe methods, to protect sensitive data from cyberthreats. Telum II enhances security on A and Z bus links, ensuring compliance and protection.4
Both Telum processors ensure reliability with redundant processor connections, enabling uninterrupted service via an alternate IO path during repairs—crucial for business continuity in mission-critical applications.
The Telum processor’s low-latency design enables fast transactions and real-time analytics. Telum II boosts cache by 40%, with 10 L2 caches and larger L2, L3 and L4, improving off-chip bandwidth and latency.5
The Telum processor powers IBM z16 for fast transactions and analytics with high performance and low latency. Telum II powers IBM z17 and is expected to reach 5.5GHz, boosting speed while efficiently managing large workloads.6
Telum’s energy-efficient enhances sustainability by reducing power use and operational costs in the IBM z16. Telum II improves this further, reducing core power by up to 15% with centralized IO and DPU design for greater efficiency.7
Scale the enterprise AI workloads of tomorrow
The IBM Spyre™ Accelerator is delivered through a 75W Gen 5x PCIe-attached AI accelerator with 128GB of LPDDR5 memory.8 Designed to handle larger, more complex AI use cases, including generative AI and multimodal LLMs, it features 32 (+2) cores, each with a 2MB scratchpad and over 55% core utilization. Scalable by card and drawer, clients can leverage this efficient solution for next-generation AI challenges.
1,2, 3, 4, 5, 6, 7 All projections for Telum II are based on pre-release hardware measurements and configurations.