Tachyum Prodigy: Expanding to a 192-Core “Paper” Superprocessor

Visit Us
Follow Me

Slovakian developer Tachyum, known for its Prodigy processors designed for high-performance computing (HPC), recently announced a significant boost in its future chip’s core count – from 128 to 192 cores, marking a 50% increase. This achievement was facilitated by the implementation of new automated design tools.

Image Source: Tachyum

First introduced in 2020, the Prodigy superprocessors combine the capabilities of CPUs, GPUs, and TPUs. They are tailored to address resource-intensive tasks in cloud computing, high-performance computing, AI systems, and machine learning.

As reported by ComputerBase, Tachyum’s advancement is not limited to core count. The company also raised the count of SerDes (serializer/deserializer) blocks from 64 to 96. These blocks play a vital role in high-speed communication, transforming data between serial and parallel interfaces in both directions. Despite these improvements, the chip’s overall footprint has only slightly increased to 600 mm² from the original 500 mm², representing a 20% growth. Theoretically, more cores could be added, increasing the footprint to 858 mm². However, such an expansion could lead to memory bandwidth constraints.

The previous chip iteration featured a 16-channel DDR5-7200 interface and supported connectivity with 32 DIMM modules. In the enhanced version, support has been extended to DDR5-6400, while the combined L2 and L3 cache size has grown from 128 to 192 megabytes.

Even in simulations with just 128 cores, Tachyum claims the Prodigy processor outperforms all leading competitors’ solutions. With 192 cores, the performance gap is anticipated to widen further.

Despite these advancements, the Prodigy processor remains a theoretical concept at this stage. Tachyum aims to showcase the functionality of its first processors through demonstrations later this year.

Author Profile

Martin Harris
I'm Martin Harris, a tech writer with extensive experience, contributing to global publications. Trained in Computer Science, I merged my technical know-how with writing, becoming a technology journalist. I've covered diverse topics like AI and consumer electronics, contributing to top tech platforms. I participate in tech events for knowledge updating. Besides writing, I enjoy reading, photography, and aim to clarify technology's complexities to readers.

You may also like...

Leave a Reply

Your email address will not be published. Required fields are marked *