A roadblock needs a bulldozer to create a path. Similarly, when progress hits a wall, companies should arm themselves with radical ideas to make that wall crumble. This is something Tachyum did a few years back when advancements in the semiconductor industry became difficult due to a performance plateau. The existing approaches were falling short of developing a way to overcome the crisis.
Tachyum came up with a solution that transformed “existing software to new computational mechanisms,” without disrupting the process. This process had its focus set on unlocking of the performance of nanometer-size devices. It also enabled the company to develop benchmarks by providing solutions that offer unprecedented performance, improved efficiency, and curbs cost. The solutions are pervading diverse strata in fields like big data analytics, deep learning, edge computing, and large-scale computing. The ripple it has created so far is inspiring moves that were unthinkable even a few years back.
The last two centuries are substantially different from the previous ones in terms of speed, and as the days progress, this speed is getting enmeshed with human lives. It is becoming a defining feature for the coming years. The same is happening in the fields of computing. With data centers evolving, the need for technologies that can provide new ways of looking at the processing of applications, that too, at a cost-effective and time-saving manner, is becoming evident.
Tachyum developed a solution for all these troubling questions through their Prodigy series.
What Is Prodigy?
Tachyum’s Prodigy boasts a unique design that can “replace the majority of existing chips provisioned in hyperscale (i.e. cloud) data centers,” and it enriches performance, lowers energy consumption, improves server utilization, and space requirements. It is first-of-its-kind processor and is a perfect emblem of cutting-edge technologies. The Universal Processor comes with a High Performance Computing (HPC), General Purpose Processors, Deep Machine Learning, Explainable AI, Bio AI, Artificial Intelligence, and other AI disciplines fused in a single chip.
Tachyum’s Prodigy facilitates “a simple programming model and environment based on a coherent multiprocessor environment.” Tachyum TPUTM (Tachyum Processing Unit) has the opportunity to expand the unique value proposition of its Tachyum Prodigy® by exploring avenues that the growth of AI chipset market for edge inference provides.
The Team
The success of a company depends much on the performance of the team. Each player is like a pillar providing strength to the premise. In Tachyum, the foundation gets bolstered by Dr. Radoslav Danilak, Rodney Mullendore, Igor Shevlyakov, and Ken Wagner.
These luminaries have a combined experience of over 100 years in the market. Radoslav has significant knowledge in designing top-notch processing systems. He has around hundred patents to his name. His earlier ventures like Skyera and SandForce won him accolades. Apart from these, his experience spans across his years in Nishan Systems, Toshiba, nVidia, and Gizmo Tech.
Rodney’s knowledge encompasses designing complex storage and networking chips and systems. He has around 28 patents under his belt and he partnered with Radoslav in his other ventures like Skyera and SandForce as well.
When it comes to compilers, system software, and tools development, Igor’s dexterity is unquestionable. His work at Skyera brought him recognition. He also spearheaded projects like the AOT Java compiler for Excelsior’s JET project.
Lastly, Ken is also known as a Co-founder of Silicon Analytics (a chip synthesis tools company), Wave Computing (a machine learning company), and Theseus Logic (a clockless logic company). His experience in the Aerospace & Defense industry is commendable where he “developed simulators for astronaut zero-G training.”
Under the leadership of these inspiring individuals, Tachyum is now all set to launch its new Prodigy chip in the first half of 2021. The year will also see the launch of “a four-socket system that will go into Open Compute Project racks.”
