San Jose-based startup Tachyum is committed to providing “universal” processors for HPC and AI computing. Posted progress For that work, we shared what the current state of the processor is. According to the company’s press release, the processor successfully booted the Linux kernel with an emulated design running on a field programmable gate array (FPGA).
Called this processor genius, Tachyum wants to create a universal solution for hyperscale, HPC, and AI purposes. The current situation requires accelerators dedicated to these types of workloads, and the solution is to use the CPU in combination with a dedicated accelerator. That’s what Tachym is trying to change on the Prodigy processor.
According to the company’s announcement today, the first Linux will boot on the hardware, albeit with a slight emulation. Tachyum booted Linux on the FPGA. So this is a strict emulation, not a real silicon implementation.
FPGAs are typically used to emulate hardware before the actual silicon arrives, so companies can prepare software to run on the actual silicon and test some preliminary performance. All digital logic companies test their designs on FPGAs before tapeting out the actual silicon.
Tachyum succeeded in “running a Linux boot, running a short user mode program, and shutting down the system with a fully functional FPGA emulation system.”
The next phase of Prodigy processor delivery is debugging and validation. Engineers are running trillions of test cases to find potential bugs in their processors. As they are resolved, the design is approaching manufacturing. A full 4-socket reference design motherboard will be available in the first quarter of 2022 when Tachyum wants to offer the Prodigy design to its first client.
https://www.tomshardware.com/uk/news/tachyum-prodigy-boots-linux Tachyum boots Linux on an emulated FPGA-based Prodigy processor