Ada Lovelace (microarchitecture)

Ada Lovelace (microarchitecture). It is a graphics processing unit (GPU) microarchitecture developed by Nvidia, succeeding the Ampere (microarchitecture) and named for the mathematician Ada Lovelace. First announced in September 2022, it powers the consumer GeForce 40 series and professional RTX 4000 series of graphics cards. The architecture introduced significant generational improvements in ray tracing performance and artificial intelligence-driven features through its new Streaming Multiprocessor design and third-generation RT Cores.

Overview

The Ada Lovelace architecture was formally unveiled by Nvidia CEO Jensen Huang during a keynote at GTC in September 2022. It is fabricated on a custom version of TSMC's 4N process node, marking a substantial shift from the Samsung 8 nm process used for its predecessor, Ampere (microarchitecture). The architecture's namesake, Ada Lovelace, is celebrated as a pioneer in computer programming. Key technological focuses for the microarchitecture include dramatically enhanced performance in real-time ray tracing and the acceleration of AI workloads, which are central to features like DLSS 3.

Architecture

The core building block of the Ada Lovelace microarchitecture is an enhanced Streaming Multiprocessor (SM) that offers increased performance per watt. A major innovation is the inclusion of third-generation RT Cores, which feature new Opacity Micromap and Displaced Micro-Mesh engines to accelerate ray tracing traversal and intersection calculations. The architecture also debuts fourth-generation Tensor Cores with the FP8 and Hopper Transformer Engine, boosting AI inference for technologies like DLSS. Furthermore, it introduced the Optical Flow Accelerator for frame generation and supports AV1 encoding via its eighth-generation NVENC encoder.

Performance

Performance claims by Nvidia positioned Ada Lovelace as offering a generational leap, particularly in ray tracing-heavy workloads. The flagship GeForce RTX 4090 was demonstrated as being capable of rendering complex ray tracing scenes at playable frame rates where previous architectures struggled. The integration of DLSS 3 Frame Generation, powered by the new AI and optical flow hardware, allowed for substantial increases in perceived frame rate in supported titles like Cyberpunk 2077 and Microsoft Flight Simulator. Independent reviews from outlets like Tom's Hardware and AnandTech generally confirmed significant rasterization and ray tracing performance gains over Ampere (microarchitecture), especially at higher resolutions.

Products

The Ada Lovelace microarchitecture first launched in the consumer GeForce 40 series, starting with the GeForce RTX 4090 in October 2022, followed by the GeForce RTX 4080 and GeForce RTX 4070 Ti. The professional visualization lineup, the RTX 4000 Ada Generation, includes workstation cards such as the RTX 6000 Ada Generation. These products are found in systems from partners like Dell Technologies, HP Inc., and Lenovo. The architecture also powers the Nvidia DGX H100 AI compute platform's graphics subsystem and the NVIDIA Jetson Orin platform for edge computing.

Reception

The reception of the Ada Lovelace architecture was largely positive for its performance but mixed regarding its market positioning and power consumption. Reviewers praised the transformative effect of DLSS 3 and the exceptional ray tracing performance in flagship cards. However, outlets like Gamers Nexus and Digital Foundry critiqued the high MSRP of several models and the significant thermal design power (TDP) of cards like the GeForce RTX 4090. The architecture was seen as solidifying Nvidia's lead in features for PC gaming and professional rendering against competitors like AMD's RDNA 3 and Intel's Arc series.

Category:Nvidia microarchitectures Category:Graphics processing units Category:2022 in computing