US-Slovak processor company in IPCEI bid
Tachyum was founded by CEO Radoslav Danilak in 2016 and the company has an engineering site in Bratislava, Slovak Republic, where Danilak was born.
The IPCEI funds, worth up to €49 million if granted, would be used to accelerate the development of a second-generation of Tachyum’s Prodigy processor in a 3nm manufacturing process.
The Prodigy 2 would be used for high-performance computing and artificial intelligence applications. It would enable one AI Zettaflop computer and more than ten data processing exaflop computers, Tachyum said.
Tachyum describes its Prodigy architecture as being the template for a “universal” processor that can run HPC applications, convolutional AI, explainable AI, general AI, bio AI, and spiking neural networks, plus normal data center workloads, using existing standard programming models.
At present – and in contrast – data centers use a combination of disparate hardware, CPU, GPU, TPU.
“As we near the completion of our first phase of delivering Prodigy to market in order to transform data centers throughout the world, we have already begun work on the next generation of Prodigy processors that will enable us to help transform economic opportunities for society at large,” said Danilak, in a statement.
Danilak had previously said that the first Prodigy chips – built in a 7nm FinFET manufacturing process – would be sampling in 1Q21 with volume production in 2H21. That appears not to have happened.
“Being selected as a key participant in this IPCEI allows us to re-imagine how we can best revolutionize industry by developing a 3nm version of Prodigy that will enable superhuman brain-scale computing. We look forward to working with the commission and contributing to the success of this project.”
Related links and articles:
News articles:
- Universal processor startup goes for split tape-out
- ‘Universal’ processor startup gains supercomputer design win
- Sparse AI chip developer raises ‘tens of millions of dollars’ in Series A
- GraphCore joins €2.6m project for real world AI
