NVLink

Headlines · 2

1. 4d
   Four Tesla V100 GPUs beat two DGX Spark units in inference test

   The YouTube channel Jun Yamog posted an Oct 12 test running four 2017-era Nvidia Tesla V100 data-center GPUs on an NVLink carrier board with a 4-bit quantized Qwen3.8-Flash-Next model. The cluster was reported 40% faster than two DGX Spark units and 230% faster than an M5 Ultra Mac Studio on long-prompt processing, and 40% and 163% faster respectively on short prompts. Short-prompt token generation reached 100.6 tok/s, about 121% faster than the two DGX Sparks, while the M5 Ultra measured 44.5 tok/s. The tester said the large gap versus Apple Silicon may partly reflect weaker software optimization.

   IT之家 AI · 🔥 0
2. 4d
   Modded 4× Tesla V100 rig beats two DGX Sparks and M5 Ultra in local LLM inference

   YouTuber Jun Yamog fitted four 2017-era NVIDIA Tesla V100 cards (32GB HBM2, 900GB/s bandwidth) onto an NVLink carrier board and used milled-down A100 heatsinks to cool them, then benchmarked local inference with a 4-bit quantized Qwen3.8-Flash-Next. The quad-V100 setup processed long prompts roughly 40% faster than two DGX Sparks and about 230% faster than a single M5 Ultra, and generated 100.6 tok/s on short prompts versus 44.5 tok/s for the M5 Ultra, about 121% faster than the DGX Spark pair. A separate eight-GPU RTX PRO 6000 configuration still led most of the runs, and Yamog noted Apple Silicon's gap may largely reflect a lack of software optimization.

   Wccftech · 🔥 0

Experience and discussion from the community

Share my NVLink experienceAsk about NVLink