<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>NVLink on Rik Kisnah - Blog</title><link>https://www.rik-kisnah.ai/tags/nvlink/</link><description>Recent content in NVLink on Rik Kisnah - Blog</description><generator>Hugo</generator><language>en</language><lastBuildDate>Tue, 15 Sep 2026 09:00:00 -0700</lastBuildDate><atom:link href="https://www.rik-kisnah.ai/tags/nvlink/feed.xml" rel="self" type="application/rss+xml"/><item><title>The 2026 GPU Field Guide: H200, B200, B300, GB200, GB300, Vera Rubin, MI3xx and Helios, From the Chip to the Rack</title><link>https://www.rik-kisnah.ai/posts/ai-gpu-field-guide-2026/</link><pubDate>Tue, 15 Sep 2026 09:00:00 -0700</pubDate><guid>https://www.rik-kisnah.ai/posts/ai-gpu-field-guide-2026/</guid><description>A blueprint for how the 2026 GPU SKUs fit together, from seven years of testing what the datasheet promises against what the data center delivers: chip design, the 8-GPU box, the NVL72 and Helios racks, watts and water, the two networks, what breaks, and the commands I run before I sign off a delivery.</description></item><item><title>NVLink versus PCIe</title><link>https://www.rik-kisnah.ai/teach/gpu-ai/nvlink-vs-pcie/</link><pubDate>Tue, 16 Apr 2019 09:00:00 -0700</pubDate><guid>https://www.rik-kisnah.ai/teach/gpu-ai/nvlink-vs-pcie/</guid><description>Two roads out of every GPU. PCIe is the public road to the CPU and the network. NVLink is the private motorway between GPUs in the same box. Know the speed of each, and know which one your traffic is on.</description></item></channel></rss>