Post reply

The message has the following error or errors that must be corrected before continuing:: Warning: this topic has not been posted in for at least 120 days.
Unless you're sure you want to reply, please consider starting a new topic.

Name
Email
Subject
Message icon

Other options

Return to this topic
Don't use smileys

Verification:

Please leave this box empty:

Shortcuts: ALT+S post or ALT+P preview

Topic summary

Posted by sfgfs

- December 01, 2021, 08:58:11

Currently AMD has the advantage of TSMC's 7N (vs Nvidia using Samsung's 8N) (-> easy 30% higher power cunsumption). AMD is expected to use TSMC's 5N (although recently news regarding the switch to Samsung also appeared), but if Nvidia is also going to 5N, then AMD may need to work on their power efficiency, or go to TSMC's 4N, but Apple is already going to use up the 4N capacity (yes, AMD is not likely to use TSMC's 4N anyway). I don't like Nvidia, but I like that they introduced Raytracing and hope they go to TSMC's 5N - competition is good for customer.

Posted by Brett Gorst

- December 01, 2021, 03:46:18

I hope they go back to TSMC process for Lovelace. Samsung nodes sucked hard...

Posted by stArkOP7

- November 30, 2021, 18:29:39

Lol 7nm Ampere? It's 8nm Samsung node!

Posted by Redaktion

- November 30, 2021, 11:07:40

According to a report by DigiTimes, factories in Taiwan are already ramping up to help NVIDIA bring out its RTX 4000 series gaming GPUs in 2022, based on 5nm lithography. Notably, the report indicates that "Hopper," rather than being a new GPU generation, will be an MCM architecture for datacenter and HPC use cases.

https://www.notebookcheck.net/DigiTimes-report-outs-a-2022-timeline-for-5nm-Lovelace-based-NVIDIA-GeForce-RTX-4090-and-4080-highlights-that-Hopper-will-be-pulling-AI-datacenter-duty-instead.582114.0.html