Urgent.News

What's breaking now, across thousands of outlets.

AI

You can grab an AI mini PC housing an Nvidia Tesla GPU with 16GB of HBM2 for just $699

900GB/s of HBM2 in a 5.6-liter box for $699: a 2017 data center GPU is suddenly a memory-shortage bargain, with a few caveats in tow.

You can grab an AI mini PC housing an Nvidia Tesla GPU with 16GB of HBM2 for just $699

The Rmatamini P99S Mini AI Workstation is a compact system that houses an Nvidia Tesla V100 with 16GB of HBM2 memory, a Core i7-10850H processor, 16GB of RAM, and a 512GB SSD. Priced at $699, the system provides a powerful computing solution for AI applications. The 16GB V100 PCIe card, which is no longer supported by Nvidia's latest CUDA version, can be found for $584 new on Amazon or as low as $410 for a refurbished unit.

Despite its age, the V100 remains relevant due to its high memory bandwidth, making it ideal for token generation in large language models. However, the limited software support and relatively modest system components mean that users should carefully consider their specific needs before investing in this system.

Written by urgent.news from TechRadar's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at techradar.com →

More in AI

From Lab to Code #4: Exploring Neural Network Efficiency with Python and PyTorch

A larger neural network can often achieve better accuracy. But does that always make it the better model? When working with neural networks, it is easy to focus on a single metric: accuracy.

  • Micro CNN designed for restricted computational resources
  • Wide CNN demonstrates accuracy gained from additional computation
  • MACs provide hardware-independent estimate of computational work

More from Monday 9 November →