Mainframe chips with everything!

If you were IBM and you were deciding where to focus for your next mainframe chip, what would you’re thinking be? There’s security, which is a massive issue everywhere. The 17^th “Cost of a Data Breach Report”, researched by the Ponemon Institute for IBM found that the average cost for a data breach rose from $3.86 million last year to $4.24 million this year. A data breach in the USA averaged at $9.05 million per incident. For breaches between 50 million and 65 million records (mega breaches), the average cost was $401 million. And the average time to detect and contain a data breach was 287 days – 212 days to detect a breach and 75 days to contain it. IBM has been producing chips focused on security for a number of years.

A second big focus for computing is the cloud, but IBM can’t produce cloud chips.

So, where is the next major focus? The answer is Artificial Intelligence (AI), and, in a way, it still links back to security.

IBM’s new chip, which is called Telum, come with a new architecture designed to handle AI workloads faster, which, as a consequence, provides improved security and fraud detection for mainframes used by financial services organizations such as banks and insurance companies.

Announced at the Hot Chips conference in August, the Telum processors contain on-chip acceleration for AI inferencing while transactions are taking place. Traditionally, data needed to be moved for inferencing to take place. However, with the Telum processors, the accelerator is positioned closer to the data and applications so users can carry out high-volume inferencing for real-time transactions without calling on off-platform AI functions. And that’s what makes the process so fast. That’s what will allow financial institutions to move from a fraud detection posture to a fraud prevention posture.

Let’s take a look at the details. IBM has already been developing the chip for three years, and it’s expected to see the light of day in the first half of next year in the new IBM mainframes that are expected to be announced then.

The chip has a centralized design, allowing users to accelerate credit approval processes and identify trades and transactions likely to fail, and enhance rules-based fraud detection, loan processing, clearing and settlement of trades, anti-money laundering, and risk analysis, according to the IBM.

The 7-nanometer chip has a dual chip module design containing 22 billion transistors and was created by the IBM Research AI Hardware Center. Samsung will manufacture the processor, which contains 8 processor cores with a deep super-scalar out-of-order instruction pipeline, running with more than 5GHz clock frequency, optimized for the demands of heterogenous enterprise class workloads.

IBM explained that: “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.”

AI workloads have higher computational requirements and operate on large quantities of data. For this to work successfully, the CPU and AI core need to be integrated on the same chip for low-latency AI inference, which is what Telum does.

“We see Telum as the next major step on a path for our processor technology, like previously the inventions of the mainframe and servers”, says a blog post by IBM’s AI Hardware Center. “The challenges facing businesses around the world keep getting more complex, and Telum will help solve these problems for years to come.”

Christian Jacobi, an IBM distinguished engineer and chief architect for IBM Z processor design, explained that the new on-chip AI accelerator achieves a 40% gain in performance at the socket level, which means every socket in the new processor, compared with processor sockets in the existing z15, performs up to 40% more work.

This additional speed is achieved by adding more cores and larger caches. Jacobi added that IBM will further optimize the firmware and software stack that is going to run on the next generation of mainframes.

So, something to look forward to next year. A chip that focuses on AI and also helps with security. And, according to Barry Baker, IBM’s vice president of product management for IBM Z & LinuxONE, as users open up their systems “to hybrid cloud, they need to modernize their application data estate. The things we’re doing down at the silicon level accelerate and are aligned with that overarching strategy. There’s a lot of work we do on a regular basis with partners to help enable that.”

Telum is a new chip that seems to tick all the boxes.