Tachyum, the “universal processor” startup, unveils the full system Prodigy emulator ahead of sampling later this year • The Register

0

Tachyum, who has promised a “universal processor” dubbed Prodigy over the past three years, said it was one more step towards realizing that commitment with the development of a motherboard for its FPGA emulator that allows customers to test a complete Prodigy system.

Trained in 2017 by Skyera and SandForce co-founder Dr. Radoslav “Rado” Danilak, Wave Computing co-founder Ken Wagner, engineer Igor Shevlyakov and hardware architect Rod Mullendore, Tachyum was straight out of the door with a bold projection for its material: a tenfold gain in performance.

Tachyum was funded on the promise of what he described as the “Cloud Chip,” which was officially announced in 2018 as Prodigy – an alternative to Intel’s server-centric Xeon processor family with its own instruction set architecture and compatibility with legacy ARM and x86 software delivered through an emulation layer, but somehow running faster than when run natively on a Xeon chip.

At the time, the new chip design was supposed to be registered and manufactured at TSMC in 2019. It is now 2021 and there is no Prodigy. Although he joined an EU project to build an exascale supercomputer, Tachyum has yet to build a commercially available silicon prodigy – although he has, he says, proven the concept on a field programmable gate array (FPGA).

Now that same FPGA prototype can be installed in a functioning computer system. “Our IO motherboard, in conjunction with our CPU motherboard, allows our engineers to fully test the functionality of Prodigy,” Danilak said of his company’s latest hardware.

“Together, these two FPGA-based boards form the basis of a system that can be cascaded to fully emulate a 128-core Prodigy processor, capable of moving the entire world into a greener era by enabling AI to the scale of the human brain. “

The FPGA prototype has been made available to “early adopters partners,” the company confirmed, with the project on track to sample the entire Prodigy system later this year. Feedback from testing with the FPGA prototypes will trickle down to an aggressive 5nm process node, Tachyum said. A four-socket motherboard, meanwhile, is expected to be produced by the end of the year.

Although it continues to work with power-hungry FPGA emulation, Tachyum sticks to its main claims: a ten-fold reduction in power consumption between core and core, the ability to beat the “GPU on the Nvidia’s “faster” in high-performance computing workloads, AI training and inference, and a fourfold decrease in total cost of ownership compared to established competitors.

It remains to be independently verified whether Prodigy will deliver on its promises and the promise of being able to switch between workloads written for radically different architectures on the fly while running them faster than natively. Interested parties may, however, request access to the emulation system on the Company Website. ®

Share.

About Author

Leave A Reply