GeForce 400 series
Nvidia graphics cards introducing the Fermi microarchitecture in 2010.
The GeForce 400 series comprises graphics processing units from Nvidia that debuted the Fermi microarchitecture. Released in March 2010 after delays, these chips brought Direct3D 11 support to consumers but faced criticism for high heat and power use compared to rivals.
- Introduced the Fermi microarchitecture in March 2010
- GF100 chip contains 3.0 billion transistors on 40 nm process
- First Nvidia consumer chips to support Direct3D 11
- Consumer cards limit double precision to one operation per four cycles
- Competed directly with ATI Radeon HD 5000 series
What it is
The GeForce 400 series represents a line of graphics processing units developed by Nvidia. These chips marked the commercial introduction of the Fermi microarchitecture, a significant shift from the previous Tesla design. The series launched in March 2010, bringing support for modern graphics standards like Direct3D 11 and OpenGL 4.0 to consumer hardware. While aimed at gamers, the underlying architecture also targeted professional computing and high performance tasks through its enhanced double precision capabilities.
How it came about
Nvidia initially planned to release this series in November 2009 but encountered delays that pushed the launch to March 2010. The company published a white paper in September 2009 detailing the new architecture, which was named after physicist Enrico Fermi. The design focused on general purpose processing, allowing thousands of threads to run simultaneously. This approach mirrored supercomputer architectures more than traditional graphics chips, aiming to simplify programming via C++ support and unified memory addressing.
Inside the chip
The flagship GF100 chip contains 3.0 billion transistors and features sixteen streaming multiprocessors. Each multiprocessor holds 32 CUDA cores, totaling 512 processors. The design includes a large 768 kB L2 cache shared across all clusters. Memory configuration is flexible, with 64 kB per cluster split between cache and shared memory. Unlike previous generations, Fermi supports a 40-bit virtual address space, enabling easier memory mapping for complex applications. Consumer cards restrict double precision performance to one operation per four cycles, while professional models allow one per cycle.
How it was made
TSMC manufactured all chips in the series using a 40 nanometre process. The largest die, the GF100, integrates 3.0 billion transistors. Smaller variants like the GF108 contain 585 million transistors. No fully enabled GF100 GPUs were sold to consumers; the GTX 480 shipped with one multiprocessor disabled due to yield or thermal constraints. The manufacturing process supported GDDR5 memory interfaces, with consumer cards typically featuring 256 MB attached to each enabled memory controller.
Where it went
These graphics cards served desktop PCs for gaming and general computing. The architecture also powered Nvidia Tesla and Quadro lines for data centres and workstations. Professional users benefited from error correcting code memory and higher double precision rates. The chips supported Visual Studio and C++, making them suitable for scientific simulations and high performance computing tasks. Consumer models provided access to Direct3D 11 features for modern games and multimedia applications.
Versions and family
The series includes entry level models like the GT 405, mid range options such as the GTX 460, and high end cards like the GTX 480. The GT 405 is an exception, using the older Tesla core instead of Fermi. Higher end consumer cards had disabled components to manage heat, whereas the later GF110 redesign enabled all cores for the GeForce 500 series. Compute capability varies, with GF100 offering version 2.0 and smaller dies providing version 2.1.
Why it matters
The series introduced key programmability features that influenced subsequent GPU designs. It established Fermi as a foundation for both gaming and professional compute markets. However, its thermal profile and power demands set a precedent for cooling challenges in high performance graphics. The move to legacy driver support in 2018 marked the end of its active lifecycle, though it remained compatible with Direct3D 12 at feature level 11_0.
Problems and reception
Users reported high temperatures and significant power consumption during operation. Performance gains over the previous generation were viewed as modest relative to the energy cost. Competitor AMD released a promotional video mocking the heat output, suggesting the chips could fry eggs. This public perception highlighted the trade offs between the new architectural features and operational efficiency in the initial Fermi products.


Questions people ask about GeForce 400 series
What microarchitecture does the GeForce 400 series use?
The series introduces the Fermi microarchitecture, named after physicist Enrico Fermi. It represents a major step forward from the previous Tesla design, focusing on general purpose computing capabilities alongside graphics processing.
Why were some components disabled on the GTX 480?
No fully enabled GF100 GPUs were sold. The GTX 480 shipped with one streaming multiprocessor disabled, while the GTX 470 had two disabled. This was likely due to manufacturing yields and thermal management constraints inherent in the initial 40 nm design.
Does the GeForce 400 series support Direct3D 12?
Yes, all products in the series support Direct3D 12.0, but only at feature level 11_0. This allows them to run newer APIs while utilizing the underlying hardware capabilities defined by the earlier Direct3D 11 specification.
How does the GT 405 differ from other cards in the series?
The GT 405 is an OEM-only card that does not use the Fermi architecture. Instead, it relies on the older GT218 Tesla core. It lacks OpenCL support and only supports DirectX 11.1 at feature level 10_1, unlike the rest of the series.
Good reading and sources
- NVIDIA GeForceNvidia, Article
- Checking your browserAnandTech, Article
- NVIDIA Fermi GF100 GPUs - Too little, too late, too hot, and too expensiveZDNET, Article
- GeForce GTX 480 And 470: From Fermi And GF100 To Actual Cards!Tom's Hardware, Article
- GeForce 400 series on WikipediaEncyclopedia
- More photos on Wikimedia CommonsPhotos
- Wikidata recordOpen data
The facts on this page come from the Wikipedia article "GeForce 400 series", written by its contributors and used under the CC BY-SA 4.0 licence, and from its Wikidata record (CC0). EveryChip rewrote the text in its own words, with AI assistance, and compares every page with the article to make sure no passage is copied. It may be reused under the same CC BY-SA 4.0 licence with a link back.
It may contain mistakes, and figures change over time. Nothing here is investment or buying advice. If something is wrong, tell us. Last refreshed 2026-11-14.
Like GeForce 400 series
All graphics processorsMore from Nvidia
All of them
WorldGeForce 500 series
Nvidia Fermi-based graphics processor family from 2010GPU · GPU architecture · 2010
🇺🇸 United StatesTILE-Gx
VLIW multicore processor family by Tilera with mesh network architectureCPU · VLIW · 2010
WorldGeForce 300 series
OEM-only rebranded Tesla GPUs from late 2009GPU · GPU architecture · 2009