With the rapid development of productive artificial intelligence technologies, Google is also taking important steps in the field of hardware to meet the high processing power, memory and communication capacity required by these technologies. At the Google I/O 2024 event, the company announced its 6th generation TPU, Trillium, the most powerful and energy-efficient Tensor Processing Unit (TPU) to date.
Trillium: A breakthrough in performance and efficiency
Trillium TPUs deliver a 4.7x increase in peak compute performance per chip compared to TPU v5e. Google has doubled the High Bandwidth Memory (HBM) capacity and bandwidth, and also doubled the inter-chip interconnect (ICI) bandwidth over TPU v5e.
Trillium comes with third-generation SparseCore, a dedicated accelerator for processing ultra-large embeddings commonly used in advanced ranking and recommendation workflows.
Trillium TPUs make it possible to train the next wave of AI models faster and deliver these models at lower latency and lower cost. The 6th generation of TPU is also the most sustainable TPU: Trillium TPUs offer more than 67% energy efficiency compared to TPU v5e.
Trillium can scale up to 256 TPUs in a single high-bandwidth, low-latency bucket. Beyond this pod-level scalability, Trillium TPUs with multislice technology and Titanium AI Processing Units (IPUs) can scale to hundreds of pods, connecting tens of thousands of chips per second with a multi-petabit data center network, creating a building-scale supercomputer.
Pioneer of AI-driven hardware
Google has been pushing the limits of scale and efficiency by developing AI-specific hardware for more than a decade. Google began developing TPU v1, the world’s first purpose-built AI accelerator, in 2013 and followed this with the first Cloud TPU in 2017.
Without TPUs, Google’s most popular services, such as real-time voice search, photo object recognition, and interactive language translation, as well as cutting-edge core models such as Gemini, Imagen, and Gemma, would not be possible. The scale and efficiency of TPUs enabled fundamental work on Transformers, which form the algorithmic foundations of modern generative artificial intelligence.
Trillium and AI Hypercomputer
Trillium TPUs are part of Google Cloud’s AI Hypercomputer, a groundbreaking supercomputer architecture designed specifically for cutting-edge AI workloads. AI Hypercomputer integrates performance-optimized infrastructure (including Trillium TPUs), open source software frameworks, and flexible consumption models.
Empowering developers with its support for open source libraries such as JAX, PyTorch/XLA and Keras 3, Google also facilitated model training and presentation by partnering with Hugging Face on Optimum-TPU.
AI Hypercomputer also offers flexible consumption models required for artificial intelligence and machine learning workloads. Dynamic Workload Scheduler (DWS) addresses access to artificial intelligence and machine learning resources and helps customers optimize their spend.
Flexible launch mode can program all the accelerators needed simultaneously, regardless of your entry point, improving the experience of bursty workloads such as training, tuning or batch jobs.
Powering the next wave of AI innovation
Trillium TPUs will power the next wave of AI models and agents, and Google is excited to bring these advanced capabilities to its customers. Autonomous vehicle company Nuro is committed to creating a better daily life through robotics by training its models with Cloud TPUs.
Deep Genomics is powering the future of drug discovery with AI and looks forward to how their next foundational model, powered by Trillium, will transform patients’ lives. Deloitte, Google Cloud’s AI Partner of the Year, will deliver Trillium to transform businesses with generative AI support.
Trillium TPUs are the result of over a decade of research and innovation and will be available later this year. Google aims to usher in a new era of artificial intelligence innovation with Trillium.
You can click the link below to watch the event.
shiftdelete.net