精选60Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)模型发布/更新
GB300 GPU 达 1648 TFLOPS,MFU 33%
1648 TFLOPS/GPU on GB300 sounds really cool, and is.
1648 TFLOPS/GPU on GB300 sounds really cool, and is. this is with MXFP8. So 33% MFU. DeepSeek themselves, in late 2023, did ≈25% on H800s (they calculate MFU based on BF16 here, but heavily used FP8). A 4.3x gap in per-GPU speed. About 2.5x in Watt/tokens? But I expected more.
更进一步:量化金融体系
看懂新闻只是起点——沿量化金融路径,把它变成能交付的工程能力