Sign up for our newsletter and get the latest HPC news and analysis.
Send me information from insideHPC:


Tachyum Joins CXL High-Speed Interconnect Consortium

Today Tachyum announced it has joined the Compute Express Link (CXL) Consortium, a group of companies dedicated to advancing a new, open-industry standard interconnect designed to address the increasing demands of high-performance computational workloads.

CXL is a high-speed CPU-to-Device and CPU-to-Memory interconnect that offers high-bandwidth, low-latency connectivity between host processors, systems and devices, such as FPGAs, intelligent networking and storage accelerators, memory buffers, storage class memory and smart I/O devices. CXL targets heterogeneous processing and memory systems across a range of high-performance computing applications by enabling coherency and memory semantics between processors and systems. This is increasingly important as processing data in Artificial Intelligence and Machine Learning requires a diverse mix of scalar, vector, matrix and spatial architectures.

Maintaining memory coherency between CPU memory and memory on attached devices, CXL will allow resource sharing for higher performance, as well as enable reduced software stack complexity and lower overall system cost. This permits users to focus on target workloads as opposed to the redundant memory management hardware in their accelerators.

Our ultra-low power Prodigy Universal Processor is aimed squarely at the hyperscale data center market,” said Dr. Radoslav Danilak, CEO of Tachyum. “With worldwide data center capacity doubling every five years, and with AI, Machine Learning, and High-Performance Computing rapidly migrating from purpose-built systems to the cloud, the importance of a fast industry-standard interconnect, which supports memory coherency, cannot be overstated. With the proliferation of 5G networks, facilitating the rapid growth of edge compute nodes or mini data centers, high speed interconnect with cross-platform memory coherency will become absolutely critical. We expect that CXL like coherent protocols will become an integral part of the PCI Express standard in the future.”

Tachyum’s Prodigy Universal Processor Chip is the smallest and fastest general purpose, 64-core processor developed to date, requiring 10x less processor power, and reducing processor cost by 3x. Prodigy will directly enable a 32-Tensor Exaflop supercomputer and allow the building of machines more powerful than the human brain by 2021, years ahead of industry expectations. Prodigy reduces data center TCO (annual total cost of ownership) by 4x, through its disruptive processor architecture and a smart compiler that has made many parts of the hardware found in typical processors redundant. Fewer transistors, fewer and shorter wires, due to a smaller, simpler core, translates into much greater speed and power efficiency for the Prodigy processor.

For the industry, a standards-based ecosystem will be a big boost. It will open the door to innovations across many areas and I look forward to building the next generation systems”, said Kiran Malwankar, VP of Systems Engineering of Tachyum.

Tachyum joins CXL founders Alibaba, Cisco, Dell EMC, Facebook, Google, Hewlett Packard Enterprise, Huawei, Intel Corporation and Microsoft to promote CXL interconnect in data center markets, working alongside leading processor manufacturers, such as AMD, ARM, and NVIDIA, as joint-CXL consortium members.

Sign up for our insideHPC Newsletter

Leave a Comment

*

Resource Links: