Ada Lovelace
Nvidia GPU microarchitecture for gaming and data centres
Ada Lovelace is a graphics processing unit microarchitecture developed by Nvidia. Announced in September 2022, it powers the GeForce RTX 40 series and workstation cards. Built on TSMC 4N process, it introduces fourth-generation Tensor Cores and improved ray tracing.
- Built on custom TSMC 4N process node
- Features 4th-gen Tensor Cores for DLSS 3
- Supports hardware AV1 encoding at 8K
- Up to 96 MB L2 cache on AD102 die
- Successor to the Ampere microarchitecture
What it is
Ada Lovelace is a graphics processing unit microarchitecture created by Nvidia. It serves as the foundation for the GeForce RTX 40 series, workstation GPUs, and data centre accelerators. Officially announced in September 2022, this design succeeds the Ampere architecture. It is named after Ada Lovelace, a pioneering mathematician. The architecture focuses on higher efficiency and advanced rendering capabilities for modern visual computing tasks.
How it came about
Nvidia revealed the architecture during a keynote in September 2022. It followed the Ampere design released in 2020. The launch included consumer graphics cards and professional workstation models. This generation aimed to improve performance per watt compared to previous nodes. It marked a shift to a custom manufacturing process designed specifically for Nvidia by TSMC.
Inside the chip
The design features streaming multiprocessors with 128 CUDA cores each. It includes third-generation Ray Tracing Cores and fourth-generation Tensor Cores. These units support AI frame generation and mixed precision math. A new Shader Execution Reordering stage improves ray tracing performance. The memory subsystem uses a large L2 cache, up to 96 MB on top dies, reducing reliance on slower video memory bandwidth.
How it was made
TSMC manufactures the chips using its custom 4N process node. This node offers higher transistor density than the Samsung process used for Ampere. The largest die, AD102, contains 76.3 billion transistors. This density increase allows for more cores and cache within similar physical areas. The process enables lower operating voltages and better power efficiency for high-performance graphics processors.
Where it went
These chips power desktop and laptop gaming PCs via the GeForce RTX 40 series. Professional users employ them in workstations for rendering and simulation. Data centres use variants like the L40 for AI inference and graphics workloads. The architecture supports modern APIs including DirectX 12 Ultimate and Vulkan 1.4. It also handles AV1 video encoding for streaming and content creation.
Versions and family
Product lines include the GeForce RTX 4060 through RTX 4090 for consumers. Workstation cards range from the RTX 2000 to the RTX 6000 Ada Generation. Data centre offerings include the L4 and L40 models. Different dies such as AD102, AD103, and AD104 scale performance across these segments. Mobile versions use GDDR6 memory for efficiency, while desktop parts often use GDDR6X.
Why it matters
The architecture delivered significant gains in ray tracing and AI-assisted rendering. Its efficient encoder supports high-quality AV1 video at lower bit rates. The large L2 cache reduced the need for wide memory buses in some configurations. It set new performance benchmarks for consumer graphics hardware upon release. The design influenced subsequent developments in GPU caching strategies.
Problems and reception
Critics noted the lack of DisplayPort 2.0 support, limiting high-refresh-rate outputs. Competitors adopted newer display standards shortly after launch. However, the performance improvements in ray tracing and DLSS 3 were widely praised. The power efficiency gains over the previous generation were considered substantial. Some users appreciated the AV1 encoding capabilities for video workflows.
Questions people ask about Ada Lovelace
Who manufactured the Ada Lovelace chips?
TSMC manufactured the Ada Lovelace chips using their custom 4N process node. This process was designed specifically for Nvidia to improve efficiency and density compared to previous generations.
What is the main improvement in ray tracing?
Ada Lovelace introduces third-generation Ray Tracing Cores and Shader Execution Reordering. These features allow for concurrent ray tracing and shading, significantly boosting performance in ray-traced workloads.
Does it support AV1 video encoding?
Yes, the architecture includes an eighth-generation NVENC encoder with hardware support for AV1. It can encode 8K video at 60 frames per second with 10-bit color depth efficiently.
Good reading and sources
- Official websiteOfficial site
- NVIDIA Delivers Quantum Leap in Performance, Introduces New Era of Neural Rendering With GeForce RTX 40 SeriesNvidia, Article
- Nvidia’s Ada Lovelace GPU generation: $1,599 for RTX 4090, $899 and up for 4080Ars Technica, Article
The facts on this page come from the Wikipedia article "Ada Lovelace (microarchitecture)", 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 Ada Lovelace
All graphics processorsMore from Nvidia
All of them
🇺🇸 United States and 1 moreApple A16 Bionic
ARM-based system on a chip for iPhone and iPadSoC · Arm · 2022