Cerebras CEO解释晶圆级架构为何比GPU快2500倍
Andrew Feldman, co-founder and CEO of Cerebras gives the best explanation of why…
Andrew Feldman, co-founder and CEO of Cerebras gives the best explanation of why Cerebras' wafer-scale architecture is 2,500X faster than a GPU during LLM inference.
Cerebras联合创始人兼CEO安德鲁·费尔德曼给出了最清晰的解释，说明为何Cerebras的晶圆级架构在LLM推理时比GPU快2500倍。
During inference, there are 2 stages: - pre-fill, where the model first processes the user's prompt, and - decode, where it generates the answer 1 token at a time in sequence.
推理过程中有两个阶段： - 预填充，模型首先处理用户的提示，以及 - 解码，按顺序逐个生成答案的token。
During that sequencial Decode phase, before each token is calculated the model weights have to be moved from memory into compute.
在这个顺序解码阶段，计算每个token之前，模型权重必须从内存移动到计算单元。
On a GPU those weights are moved from HBM, while Cerebras keeps them in much faster SRAM spread across its very large wafer-scale processor, so the memory-to-compute movement that must happen for every token is about 2,500× faster
在GPU上，这些权重从HBM移动，而Cerebras则将其保存在遍布其超大晶圆级处理器的更快SRAM中，因此每个token所需的从内存到计算的移动速度大约快2500倍。
From The MAD Podcast with Matt Turck and Cerebras YouTube channel, (full video link in comment)
来源：The MAD Podcast与Matt Turck及Cerebras YouTube频道（完整视频链接见评论）
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