IBM Telum On-Chip AI Acceleration Processor

IBM unveils new on-chip AI acceleration processor for real-time fraud detection

To enable organizations to gain real-time insights through real-time inference, technology giant IBM developed the on-chip accelerated artificial intelligence (AI) processor. By leveraging deep learning inference to enterprise workloads, IT teams will now be able to address fraud, among other use cases, while it was taking place.

IBM Telum is designed to simplify applications and enterprise AI workloads. With the accelerator built within the vicinity of mission-critical data and applications, enterprises now have the ability to conduct high-volume inferencing for real-time sensitive transactions.

“Clients can also build and train AI models off-platform, deploy and infer on a Telum-enabled IBM system for analysis,” IBM said.

IBM creates first 2 nanometer chip technology

IBM further advances innovations on hybrid cloud, AI

Fraud and other security issues were usually discovered when they have already occurred. IBM Telum was built to mitigate or prevent or catch fraud as it occurs. Due to latency requirements, complex fraud detection often cannot be completed in real-time, according to IBM, meaning a bad actor could have already successfully purchased goods with a stolen credit card before the retailer is aware fraud has taken place.

IBM Telum design

IBM developed the new on-chip hardware acceleration for three years and it will be made available first half of 2022 for use in the banking, finance, trading, insurance applications, and customer interactions sectors.

The new chip features an innovative centralized design, which allows clients to leverage the full power of the AI processor for AI-specific workloads, making it ideal for financial services workloads like fraud detection, loan processing, clearing and settlement of trades, anti-money laundering, and risk analysis.

“With these new innovations, clients will be positioned to enhance existing rules-based fraud detection or use machine learning, accelerate credit approval processes, improve customer service and profitability, identify which trades or transactions may fail, and propose solutions to create a more efficient settlement process,” IBM said.

IBM chip-interconnection infrastructure

The chip contains 8 processor cores with a deep superscalar out-of-order instruction pipeline, running with more than 5GHz clock frequency, optimized for the demands of heterogeneous enterprise-class workloads. The completely redesigned cache and chip-interconnection infrastructure provides 32MB cache per core and can scale to 32 Telum chips. The dual-chip module design contains 22 billion transistors and 19 miles of wire on 17 metal layers.

Telum is the first IBM chip with technology created by the IBM Research AI Hardware Center. In addition, Samsung is IBM’s technology development partner for the Telum processor, developed in 7nm EUV technology node.