Introduction

NVIDIA kicked off this year’s next-generation product launch cycle with the GeForce RTX 30 Series “Ampere” graphics card announcement on September 1. We finally have our hands on a GeForce RTX 3080 Founders Edition graphics card, which is the de facto reference design for the RTX 3080. Since its 2015 “Maxwell” generation, NVIDIA sought to release the reference-design products as premium “Founders Edition” SKUs, so they form a benchmark for custom-design board partners to reach rather than baselines for them to build upon.

If you’ve read our Ampere Architecture article, you’ll know the GeForce RTX 3080 is based on a variant of the 8 nm “GA102” silicon, with 8,704 CUDA cores, 68 2nd generation RT cores, 272 3rd generation Tensor cores, 272 TMUs, and 96 ROPs. It features 10 GB of fast 19 Gbps GDDR6X memory, a technology NVIDIA co-developed with Micron, using a 320-bit wide memory interface. This is also NVIDIA’s first GeForce GPU to take advantage of the PCI-Express 4.0 x16 bus.

But the silicon is only half the story. The elephant in the room has to be the unique new board design by NVIDIA. The RTX 3080 Founders Edition is as long and tall as most high-end reference-design graphics cards, but the cooler is designed to tame a GPU with 320 W typical board power without keeping your neighbors up at night. NVIDIA did not resort to ugly liquid-cooling solutions, but rather re-imagined the cooler, which we’ll try to talk about in a bit.

In this article, we unbox the GeForce RTX 3080 Founders Edition graphics card and bring you detailed pictures of the card exterior. As of today, NVIDIA only allows “photography of the card, no disassembly, no plugging in, no installing, no powering on, no performance results or hints of any kind.” Our comprehensive GeForce RTX 3080 Founders Edition per