Processor manufacturer Tachyum claims to have demonstrated the execution of a non-native application under Linux running on the emulation system for its Prodigy FPGA, which is the basis for its patented Prodigy universal processor.
This is a major milestone in Tachyum’s mission to demonstrate that its universal processor can handle the best CPUs And best GPUsprior to moving to general availability in 2024.
The chip’s architecture does not include dedicated hardware for running x86 applications, which are responsible for the broadest data center installation base, as this hinders performance. But customers with Tachyum hardware can, according to the demo, emulate such applications.
“Demonstrating the ability to run binary x86-64 applications on the Prodigy processor emulation is an important milestone for Tachyum and further validates our architecture before tape release,” said Tachyum Founder and CEO Dr. Radoslav Danilak.
“Proving all of these existing binary applications on our FPGA emulator is critical to getting Prodigy ready for volume production next year, allowing us to start fulfilling the billions of dollars of orders we already have in our sales pipeline.”
Flip my chip up
This comes just weeks after the company announced an increase in the core count of its Prodigy Universal Processor to a monstrous 192.
Following a technique known as register transfer level (RTL) design coding, Tachyum began work on completing the physical design for Prodigy. With a new set of data analysis tools, the team was able to increase the number of cores by 50%.
Other improvements included increasing the L2/L3 cache from 128MB to 192MB, support for DDR5 7200 RAM in addition to DDR5 6400 memory, as well as a larger package. This can accommodate an additional 32 serial connections and as many as 32 DIMMs connected to a single Prodigy chip.
The Intel rival, based in Slovakia and the US, wants its proprietary Prodigy processor to run conventional workloads and AI apps on the same chip.
This means that business customers can abandon the need for special and specialized AI-centric chips from companies, for example AMD and Nvidia in addition to the running infrastructure equipped with Intel’s x86 CPUs. The result, according to Tachyum, is much lower operating and capital expenditures, in addition to better server utilization.
As a general-purpose processor, the architecture allows Prodigy to switch between conventional workloads and AI workloads, delivering high performance in both model training and inference, the company claims.
Existing customers can use the dynamic binary translator to transition legacy x86 applications to the Prodigy native instruction set architecture (ISA) within 18 months.
Until software environments are fully native, Tachyum claims that Prodigy will deliver only a twofold performance improvement, compared to the expected threefold performance improvement, compared to Intel’s fastest CPUs.