Enabling new architectures for more efficient AI data centres

SiTime Corporation has announced the differential ended SiT5977 Super-TCXO, the newest SiTime Elite RF family member. This is the only single-chip timing solution that provides the most resilient performance for AI compute-nodes with high bandwidth and network synchronisation. Applications include Smart NICs, acceleration cards, switches, and compute nodes in the data centre infrastructure market.

“Improving AI workload efficiency to reduce energy consumption and carbon emissions is an industry-wide challenge. Precision timing is one of the approaches to help solve this problem,” said Dave Altavilla, president and principal analyst at HotTech Vision and Analysis. “SiTime is the only semiconductor company fully dedicated to developing innovative timing solutions required for the complex scaling of today’s AI data centres.”

An AI cluster’s efficiency results in lower TCO and energy consumption. Efficient clusters need high-bandwidth interconnects and tightly synchronised orchestration to minimise AI accelerator idle time. Tighter synchronisation also delivers precise network telemetry for AI operations. This network telemetry improves performance and reliability by identifying and mitigating underperforming hardware before it fails.

The device replaces multiple timing components. It optimises the efficiency of AI compute clusters with 3X tighter synchronisation for data transmission and communications in real-world environments. Also, its 4X smaller size, compared to competing architectures, allows larger processors in compact systems. This enables system architects to fully leverage high-speed 800G bandwidth network connectivity and maximise utilisation of the AI cluster.

“AI training and inference are fundamentally distributed computing applications, which require accurate timing to synchronise activities,” said Piyush Sevalia, executive vice president of marketing at SiTime. “By enabling a new, more efficient architecture, the Elite RF timing solution uniquely supports more efficient AI workload processing, which may lead to higher revenue and lower TCO for data centres.”

The SiT5977 Super-TCXO streamlines AI system architecture. This new timing chip is an ultra-stable, low-jitter TCXO with the industry’s highest frequency differential output. This chip further integrates digital control to supply additional system-level programmability.

Key features include :

Environmentally robust with ±1 ppb/C frequency slope (dF/dT) for optimum performance under airflow and thermal shock.

Capable of driving 800G and higher links via 80 fs phase jitter and LVDS outputs.

Enables embedded control loops with precise digital tuning of output frequency (DCTCXO), ±400 ppm pull range and 0.05 ppt (5e-14) resolution via I2C/SPI.

Eradicates link flaps from quartz timing activity dips or micro jumps.

Resistant to shock, vibration and board bending.

Eradicates external LDOs via on-chip voltage regulators.

156.25MHz output frequency, enabling high-speed SerDes and 800G links.

The SiT5977 Super-TCXO is in production and samples are available now.