NVIDIA Reportedly Delays Next-Gen, Multi-Chip “Hopper” Architecture for 5 Nm “Lovelace”

The FPS Review may receive a commission if you purchase something after clicking a link in this article.

Image: NVIDIA

Prolific leaker kopite7kimi recently revealed that NVIDIA has delayed its oft-whispered Hopper architecture, which replaces the traditional monolithic design for a GPU that leverages multi-chip modules. Now, a user by the name of WildCracks has chimed in with the possibility that Ampere’s successor will be a new architecture named after popular English mathematician Ada Lovelace, who’s best known for her writings on the Analytical Engine.

kopite7kimi seems confident that the Lovelace architecture will be manufactured on a 5 nm process, but it isn’t clear at this point whether TSMC or Samsung will be tasked to do the job. We recently learned that NVIDIA’s relationship with the latter is still going strong, as it was just awarded a second contract to produce additional Ampere cards.

A DigiTimes report from earlier this year claimed that NVIDIA had already made a deal with TSMC to produce its multi-chip-module Hopper GPUs, but the delay obviously changes things a bit. “For next-generation GPUs based on the Hopper architecture, NVIDIA has already pre-booked TSMC’s 5 nm production capacity in 2021, and is also in discussion with Samsung for smaller volume orders,” reads a translation provided by RetiredEngineer.

Tsing Mui
News poster at The FPS Review.

Recent News