GPU manufacturer NVIDIA was attacked by a hacker organization ransomware virus at the beginning of the year.
After the internal system was compromised, more than 1TB of data was leaked including drivers, original design files, and firmware, and a large number of exposure images included unpublished GPUs, including Ada Lovelace, Hopper, and Blackwell architectures.
In the past few months, NVIDIA released the Hopper architecture and Ada Lovelace architecture GPUs, which are respectively aimed at data centers and game graphics cards.
According to the cycle of NVIDIA’s technological change, the next time will be about two years later.
The successor to the Hopper architecture is the Blackwell architecture, which is known to have two GPUs, GB100 and GB102.
Foreign media Wccftech reported that at the Arete Technology Conference a few days ago, NVIDIA Vice President and Accelerated Computing Director Ian Buck reiterated NVIDIA’s commitment to updating the main GPGPU architecture plan every two years, confirming that the Blackwell architecture GPU will be launched in 2024.
Currently, the Hopper architecture GPU uses TSMC’s 4nm process, and it is also the world’s first GPU using HBM3.
Huida’s next-generation Blackwell architecture GPU will continue to cooperate with TSMC and adopt a 3-nanometer process.
The name of the Blackwell architecture is rumored to be named after the late American statistician and mathematician David Harold Blackwell (1919~2010), who received a bachelor’s, master’s and doctorate in mathematics from the University of Illinois at Urbana-Champaign in 1938, 1939 and 1941.
He became a full-time professor at Howard University in 1947, entered the University of California, Berkeley until retirement in 1954, and became the first black tenured professor in school history, and was elected the first black academician of the American Academy of Sciences in 1965.