在过去的10多年，半导体集成电路器件密度，功率效率和成本改进的进度已经大大放缓。在业绩停滞危机的阴影下，半导体行业正在讨论摩尔定律的消亡的问题。这个危机的根本原因在于现有的硬件，软件和系统 设计方法不在能够有效的克服纳米级器件的物理上的根本限制。我们需要一种全新的解决方案，顺畅地将现有软件，算法转换为新的计算机制 ，从而释放出纳米级半导体器件的性能潜力。Tachyum的创始人拥有解决半导体器件物理机制所带来限制问题的成功经验。Tachyum将发挥纳米级半导体器件前所未有的性能，能效和成本优势，将具有突破性功能 的将产品推向市场，用以解决大数据分析中最复杂的问题 – 包括深度学习，移动和大规模计算。Tachyum的解决方案将会给广泛的市场和行业带来深刻的影响。
The World’s First Universal Processor for Hyperscale, HPC, and AI Workloads
Who We Are
Tachyum has developed the world’s first Universal Processor that combines the functions of a CPU, GPU, and TPU into a single homogeneous processor architecture that is faster, 10x lower power, and 1/3 the cost of competing products.
The Tachyum solution delivers never-before seen performance and efficiency to a wide range of applications including Hyperscale, HPC, and AI.
Tachyum’s revolutionary new processor maximizes performance without the need for costly and power-hungry hardware accelerators, enabling a new level of data center agility that dramatically increases performance and efficiency while lowering TCO.
How We Are Changing The World
Tachyum products significantly reduce data center power consumption, which lowers carbon emissions and helps to address climate change and keep our planet green.
In addition, Tachyum-powered supercomputers are poised to address some of the world’s most important and challenging problems, including a wide array of global health challenges, as well as the ability to provide real-time language translation from any language to any other language, making the world a true global village.
Prodigy will drive worldwide economic growth by allowing massive growth in AI. According to PWC, AI will increase global GDP by 14%, or $15.7 trillion, by 20301, and Prodigy will be a key enabler of that growth.
PwC Global AI Study: Exploiting the AI Revolution ↩︎
Multiple Product SKUs Address a Wide Range of Markets and Applications
Hyperscale Data Centers
High Performance Computing
Tachyum’s Prodigy family of high-performance processors includes a range of product SKUs that address a wide array of markets, applications, and workloads. The cutting-edge family of Prodigy solutions range from 128 cores at the high end to a 64-core volume SKU and an entry-level 32-core SKU.
By combining the functions of a CPU, GPU, and TPU, Prodigy-based servers provide unprecedented agility and flexibility, and are easily repurposed between workloads regardless of the type or complexity of the workload, while maintaining a cost and power-efficient homogeneous environment. For example, Prodigy-based servers are easily repurposed from integer workloads such as web hosting/serving to AI/ML workloads requiring high-performance vector operations.
Prodigy has a full suite of development tools, models, and libraries, including software emulators and an FPGA-based hardware emulator.
- Big AI,
- Big Data,
- Exascale Supercomputers
- Hyperscale Datacenters,
- Data Analytics
- Hyperscale Datacenters,
- Web Hosting,
- Edge Computing