Issue link: https://iconnect007.uberflip.com/i/1483130
76 SMT007 MAGAZINE I NOVEMBER 2022 Dale McHerron is senior manager at IBM Research and oversees heterogenous integra- tion research. In his presentation, Dale spoke about the need to architect the next generation of AI systems, and what it means to the infra- structure to run AI systems at this high level of technology. Nolan Johnson: Dale, you just finished your pre- sentation and I'd love to get a quick summary. Dale McHerron: Sure. My focus is on high perfor- mance computing, with a special interest in AI and AI workloads, and how we need to archi- tect next generation systems to support what we see happening in AI. I talked about where we are today from a traditional IBM high per- formance computing standpoint, and some of the aspects of our z Systems® (zHPF) main- frame. en I started talking about the emerg- ing confluence of the need for much more com- pute power, while simultaneously silicon scaling is starting to slow down. ere is the need for architecting for new workloads, which drives a huge amount of memory. How do you bring all that together? at all has an impact on what we will need going forward from the packaging. Johnson: You had an interesting statistic that AI compute power requirements are doubling every three and a half months. Can you talk about that? McHerron: e algorithms are getting so sophis- ticated. As you can see, AI is raising some eye- brows. People are really doing things with it. ose algorithms are getting so complex, and so computationally complex in terms of train- ing, which is much more complex than infer- encing. Training is how you train the model; inferencing is actually using it in the real world. You look at the training requirements for these very sophisticated algorithms, and they need to double the amount of compute every three and a half months because they're getting so complex. Our friends in the soware and algo- rithm worlds are really making a lot of leaps and gains in terms of the complexity and the ability of these new algorithms. Johnson: at just feels like a pace well more than what you'd expect from Moore's Law. at's driving the hardware even harder than we traditionally think—even at a time where some people are saying Moore's Law is broken. McHerron: It is. ere's been a lot in the indus- try lately about how much energy data cen- ters use. I read that something like 3% of the The Critical Nature of High-performance Computing Dale McHerron