Hardware & Inference RSS

Baidu Intelligence Cloud unveils Tianchi supernode at WAIC to expand scaling boundaries of domestic computing power

At the World Artificial Intelligence Conference (WAIC), Baidu Intelligence Cloud announced the Tianchi supernode, a new computing infrastructure designed to push the limits of scaling for China's homegrown computing power. The supernode aims to support large-scale AI model training and inference, enhancing the capabilities of domestic chips and systems.

BaiduBaidu
Baidu ERNIE (GNews)24.07 · 11:03

Mini-PC on Strix Halo under parallel load: 236 tok/s on 32 concurrent requests and three errors

A Beelink GTR9 Pro with Ryzen AI Max+ 395 achieved 236 tok/s aggregate on 32 concurrent requests in short runs, sustaining 226 tok/s average over 30 minutes without throttling. The author discovered a reproducible throughput drop between 8 and 10 concurrent clients across all tested models, and found that speculative decoding (MTP) accelerated single-client performance by 42% but became a penalty under concurrency.

Google/DeepMindGoogle/DeepMind
Habr — хаб NLP24.07 · 10:04
Fresh news