Aug 4, 2021 · Artificial Intelligence

How We Accelerated Feature Hashing for Ad Ranking on GPUs

This article explains how Tencent's Light platform reduced the massive overhead of feature hashing in ad‑ranking by moving integer‑to‑string conversion and hash computation to the GPU, introducing custom contiguous string tensors, and achieving up to 12× speed‑up on V100 GPUs.

GPU OptimizationTensorFlowad ranking

0 likes · 14 min read

How We Accelerated Feature Hashing for Ad Ranking on GPUs

Alibaba Cloud Developer

Dec 14, 2017 · Artificial Intelligence

How Alibaba’s eXtreme Parameter Server Powers Billion‑Scale Machine Learning

Alibaba’s eXtreme Parameter Server (XPS) platform tackles the challenges of training models on billions of samples and trillions of features by employing streaming learning, feature hashing, dynamic sparsity, asynchronous checkpointing, and high‑performance communication, enabling efficient, fault‑tolerant distributed AI at massive scale.

Alibabadistributed machine learningfeature hashing

0 likes · 21 min read

How Alibaba’s eXtreme Parameter Server Powers Billion‑Scale Machine Learning