Nvidia’s GPU Names Explained: Ampere, Hopper, Blackwell, Rubin, Feynman

GPU Roadmap (2024‑2028)

2024 – Release of the Blackwell GPU family.

Early 2025 – Full shipment of Blackwell GPUs.

Mid‑2025 – Announcement of Blackwell Ultra.

Late 2026 – Launch of the Vera Rubin GPU platform ("super‑chip").

Late 2027 – Announcement of Rubin Ultra.

2028 – Introduction of the Feynman architecture (details pending).

Architectures and Their Namesakes

Ampere (2020)

Named after French physicist André‑Marie Ampère (1775‑1836), a founder of classical electromagnetism. The Ampere architecture introduced the A100, H100 and H800 GPUs; the leading “A” denotes the generation while the following numbers indicate performance tiers.

Hopper (2022)

Honors computer scientist Grace Hopper (1906‑1992), a pioneer of compiler technology and early programming. Hopper GPUs (e.g., H100) power the generative‑AI era; each chip costs roughly $40,000 and is widely adopted for large‑scale AI workloads.

Blackwell (2024‑2025)

Named for statistician David Blackwell (1919‑2010), co‑developer of the Rao‑Blackwell theorem, a cornerstone of statistical estimation. Blackwell GPUs target data‑intensive AI and complex systems, representing Nvidia’s most advanced chips to date. An Ultra variant is slated for mid‑2025.

Rubin (2026‑2027)

Honors astronomer Vera Rubin (1928‑2016), whose observations provided strong evidence for dark matter. The Rubin architecture is positioned as a "super‑chip" platform that will push computational boundaries, with an Ultra version expected in late 2027.

Feynman (2028)

Named after theoretical physicist Richard Feynman (1918‑1988), Nobel laureate for contributions to quantum electrodynamics. The upcoming Feynman architecture is expected to continue Nvidia’s tradition of aligning GPU innovations with pioneering scientific concepts, though detailed specifications remain undisclosed.

Earlier Naming Tradition

Prior GPU families were named after historic scientists:

Kepler – Johannes Kepler (1571‑1630), whose laws of planetary motion underpinned Newtonian physics.

Turing – Alan Turing (1912‑1954), founder of modern computer science and artificial intelligence.

Lovelace – Ada Lovelace (1815‑1852), recognized as the world’s first programmer.

Significance of the Naming Scheme

The choice of each scientist reflects a core innovation in the corresponding GPU architecture:

Electromagnetism (Ampère) ↔ high‑throughput compute.

Compiler and universal computing concepts (Hopper) ↔ generative‑AI acceleration.

Statistical estimation theory (Blackwell) ↔ data‑intensive AI workloads.

Dark‑matter research (Rubin) ↔ breakthroughs in raw computational power.

Quantum electrodynamics (Feynman) ↔ future directions in hardware acceleration.

References

https://www.businessinsider.com/nvidia-chips-gpu-names

https://en.wikipedia.org/wiki/Wiki

Code example

收
藏
，
分
享
、
在
看
，
给
个
三
连
击呗！