Covering Scientific & Technical AI | Wednesday, October 9, 2024

Tachyum Joins UCIe Community to Drive In-Package Integration 

LAS VEGAS, April 26, 2022 – Tachyum today announced it has signed a letter of agreement to become a member of the Universal Chiplet Interconnect Express (UCIe) community to further develop and support the chiplet ecosystem and drive on-package innovations.

UCIe is an open specification that defines the interconnect between chiplets within a package, enabling an open chiplet ecosystem and ubiquitous interconnect at the package level. Members of UCIe represent a wide range of industry expertise and include leaders in semiconductors, packaging, IP suppliers, foundries, and cloud services.

The UCIe 1.0 Specification is an open industry standard developed to establish a ubiquitous interconnect at the package level and covers the die-to-die I/O physical layer, Die-to-Die protocols, and software stack which leverage the well-established PCI Express (PCIe) and Compute Express Link (CXL) industry standards. The specification is available by request.

Tachyum plans to use the mature chip-to-chip technology for its Prodigy 2 processor at the 3nm level. Multi-die connectivity will play an essential role in both significantly increasing performance of Prodigy 2, as well as increasing the range of performance points and applications Prodigy 2 would be used for.

“I have dedicated much of my career to helping overcome the limitations of Moore’s Law by developing technologies that allow for smaller and more power-efficient designs,” said Dr. Radoslav Danilak, founder and CEO of Tachyum. “By joining UCIe and having Prodigy be part of its emerging chiplet ecosystem, we are creating a system that allows the mixing and matching of the best, proven components from multiple vendors to speed up adoption. We are excited to be among the UCIe members driving integration within the package.”

Prodigy has the potential to create unrivaled computational speed and vast energy saving capabilities for hyperscale, OEM, telecommunication, private cloud and government markets. Prodigy’s 10x lower processor core power consumption will dramatically cut carbon emissions associated with data center usage. Prodigy’s 3x lower cost (at equivalent performance) will also translate to billions of dollars in annual savings to hyperscalers like Google, Facebook, Amazon and Alibaba.

Tachyum’s Prodigy processor can run HPC applications, convolutional AI, explainable AI, general AI, bio AI, and spiking neural networks, plus normal data center workloads, on a single homogeneous processor platform, using existing standard programming models. Without Prodigy, hyperscale data centers must use a combination of disparate CPU, GPU and TPU hardware, for these different workloads, creating inefficiency, expense, and the complexity of separate supply and maintenance infrastructures. Using specific hardware dedicated to each type of workload (e.g. data center, AI, HPC), results in underutilization of hardware resources, and more challenging programming, support, and maintenance. Prodigy’s ability to seamlessly switch among these various workloads dramatically changes the competitive landscape and the economics of data centers.

About Tachyum

Tachyum is transforming AI, HPC, public and private cloud data center markets with Prodigy, the world’s first Universal Processor that delivers industry-leading performance, cost, and power efficiency for both specialty and general-purpose computing. When Prodigy processors are provisioned in a hyperscale data center, they enable all AI, HPC, and general-purpose applications to run on one hardware infrastructure, saving companies billions of dollars per year. A fully functional Prodigy emulation system is currently available to select customers and partners for early testing and software development. With data centers currently consuming over 3% of the planet’s electricity, predicted to be 10% by 2025, the ultra-low power Prodigy Universal Processor is critical, if we want to continue doubling worldwide data center capacity every four years. Tachyum, Co-founded by Dr. Radoslav Danilak with its flagship product Prodigy, is marching towards tape out and chip sampling in 2022, with software emulations and an FPGA-based emulator running native Linux available to early adopters. The company is building the world’s fastest 64 AI exaflops supercomputer in 2022 in the EU with Prodigy chips. Tachyum has offices in the United States and Slovakia. For more information, visit https://www.tachyum.com.


Source: Tachyum

About the author: Tiffany Trader

With over a decade’s experience covering the HPC space, Tiffany Trader is one of the preeminent voices reporting on advanced scale computing today.

AIwire