Meet Dr. Radoslav Danilak- The Man Behind Tachyum

Tachyum- The first global processor platform that offers 4x lower data center TCO and provides multiple exaflops of AI, CAPEX free

The microprocessor and GPU sector was worth USD 83.1 billion in 2019 and is expected to rise to USD 112.7 billion by 2025, with a compound annual growth rate (CAGR) of 7.3 percent from 2020 to 2025. 

Thanks to a number of factors, including the consumer electronics demand and increased acceptance of Internet of Things (IoT)-enabled products and appliances, the industry has a bright future. Furthermore, during the COVID-19 pandemic, cloud-based platforms and server environments played a key role in propelling the microprocessor and GPU market.

In-house GPUs can be used on desktops, workstations, and businesses, resulting in a diverse range of products and services for end-users. The growing use of supercomputers to speed up drug development, weather prediction, science discovery, and simulation programs has fueled the growth of in-house GPU solutions. 

On these fronts, we bring you the story of Tachyum, a firm that is enabling human brain-scale AI and advancing the entire world to a greener era, by delivering the world’s first universal processor. 

Tachyum is named after the Greek word “tachy,” which means “speed,” paired with the suffix “-um,” which is intended to imply an element from the periodic table Together it is supposed to connote the concept of the element of speed. 

Tachyum Inc. new “Prodigy” processor series, which incorporates the benefits of CPUs, GP-GPUs, and advanced AI chips in a single universal processor platform capable of running the world’s most complex computing tasks with up to ten times the processing power per watt (core vs core). Prodigy will allow a super-computational infrastructure for a real-time full capacity human brain simulation in a neural network, thanks to its groundbreaking architecture.

Speaking about its groundbreaking architecture Dr. Radoslav Danilak, CO-Founder and CEO adds, “Tachyum was established with the aim of developing the world’s first Universal Processor. We recognized that the processor industry had reached a speed ceiling and planned to develop modern technologies that would not only be faster than everything else on the planet, but also the most environmentally efficient processor ever made. At Tachyum, we have some of the brightest visionaries of our time on our core team. Our advisors have a long history in the semiconductor business. Our board of advisors includes Professor Steve Furber who created the ARM microprocessor, and Fred Weber the decade-long former CTO for AMD, and Dr. Kurt Lauk the former CEO and CFO of Audi.” 

Meet Dr. Radoslav Danilak- The Man Behind Tachyum

Radoslav grew up in the Slovak Republic. He was hired by an American Tier-1 VC to work in a start-up after a stint teaching compiler theory at a university in Slovakia. Later, as a processor architect at Toshiba, he worked on the Playstation 2 Emotion Engine processor. Radoslav architected core features of Nishan Systems’ award-winning nForce 4 and nForce 4 for Intel version chipsets, as well as a single chip network processor that replaced 20 MMC processor chips. He was one of the first architects on NVIDIA’s Tesla and Fermi GPU architectures, and a key contributor to the company’s commercial growth.

“I ventured out on my own as an entrepreneur, co-founding SandForce, and designing a disruptive Flash Memory Controller chip. The SandForce chip ushered in the era of much less expensive Flash memory (MLC) that directly enabled Apple’s MacBook Air product, which used my SandForce chip to enable the use of much less expensive multi-level cell flash (as opposed to very expensive single-level cell flash). After a successful SandForce exit (sold to LSI Corporation, San Jose, California, in 2011 for $377M), I co-founded Skyera, which produced high-density flash storage systems, sold primarily to the U.S. Military and Intelligence agencies for operational use. After another successful exit from Skyera (sold to HGST, Western Digital brand, San Jose, California, in 2014), I co-founded Tachyum Inc. in 2016, in order to solve the processor performance plateau challenge.” 

The Future- Switching to Prodigy

Tachyum’s universal processor combines the ease of programming a CPU with the speed and reliability of a GP-GPU on AI workloads, to create a universal-purpose processor capable of handling hyper-scale workloads, AI, HPC, and other challenging applications. A traditional hyper-scale data center using Prodigy-enabled servers can reduce annual TCO by a factor of 4 and provides multiple exaflops of AI during off peak hours by using idle servers which are already bought and paid for. 

“If the world’s hyper-scale data centers switched to Prodigy powered servers, it would reduce carbon dioxide emissions by more than 600 million tons per year – more than the entire airline industry. Prodigy represents a green revolution in computing.”

Dr. Danilak explains “Prodigy is the world’s first universal processor. Its homogeneous architecture delivers class-leading performance in data center, AI, and HPC workloads. Its highly efficient design, coupled with its low sell price, reduces data center TCO (annual total cost of operations) by a factor of 4x, saving hyper-scale operators billions per year. Idle servers, already bought and paid for, can be repurposed as AI training and/or inference systems during data center off-peak hours, providing an additional revenue stream, CAPEX free.” 

Looking Forward

Prodigy’s intense computing muscle can help AI use cases in particular. Convolutional neural networks, deep learning AI, symbolic AI, general AI, and bio AI are the disciplines that have emerged in recent years, each with its own set of algorithms and computing criteria. Human brain modeling, meanwhile, is in high demand by R&D ventures due to its promise of extracting information from large data sets. The Tachyum Universal Processor Platform is suitable for projects like the real-time Human Brain Project, which needs more than 10¹⁹ Flops (10,000,000,000,000,000,000 floating-point operations per second – 10 exaflops), as well as supplying computing resources for science and engineering solutions that today’s systems cannot offer.

“We have a couple of goals for our company. The first is to create the world’s fastest supercomputer that can produce the full simulation of the human brain. Our second goal is to get the technology into every datacenter worldwide so that we can help reduce the amount of electricity being consumed.” shares the CEO. 

Tachyum on Social Media

Facebook | Linkedin | Twitter