Baidu releases Tianchi 1024 Card Super Node
The story so far
Baidu releases Tianchi 1024 Card Super Node, capable of supporting 5 trillion parameter model training, to be launched next year. The node is equipped with optical engine technology, increasing inter-card interconnect bandwidth by 3 times, training performance by 2.5 times, and inference performance by 3 times. It is paired with the G3.5 KV Cache GPU Accelerator Cabinet, doubling long-context inference performance, reducing TTFT by 90% in multi-round dialogue caching scenarios, and increasing throughput by 7 times. A single node has the capability to scale to ten thousand cards.
AI-written from the reports · Updated 3 days ago
Coverage timeline
Follow the coverage to see every side of the story.
- ITHome · AIBaidu Tianchi 1024 Card Super Node Released: Can support training of 5 trillion parameter models, to be launched next year.
Baidu Tianchi's 1024 card super-node is launched, equipped with optical engine technology, inter-card interconnection bandwidth increased 3 times, training performance increased 2.5 times, and inference performance increased 3 times. Accompanied by G3.5 KV Cache GPU memory acceleration cabinet, long context inference performance increased 2 times, multi-round dialogue cache hit scenario TTFT decreased by 90%, and throughput increased 7 times. A single node can support the training of large models with 500 billion parameters, with the ability to scale to ten thousand cards, and will be launched next year.
Heat over time
- Now
- 3
- Peak
- 1012 Nov, 13:00
- 24h change
- -50%
Hover or tap the chart to see each hour; use the arrow keys on a keyboard.