NVIDIA GeForce RTX 4090 Ti rumored to have 18,176 cores and 24GB/24Gbps memory

Please note that this message is marked with a flag Common.

A rumor has surfaced regarding NVIDIA RTX 4090 Ti specifications

Quick update on RTX 4090 Ti specs shared by Kopite7kimi himself.

It looks like we’re getting closer to NVIDIA releasing faster RTX 40 graphics cards. Just a couple of days ago, a new leak confirmed that NVIDIA was working on a TITAN 800W GPU, optionally equipped with 48 GB of memory. Today we learn that the TITAN and the RTX 4090 Ti must actually be two different models following all.

According to Kopite, the RTX 4090 Ti will feature an AD102 mini GPU with 18,176 cores (1,792 more than the RTX 4090), memory upgraded to 24Gbps units, and 600W TBP. The last part isn’t entirely clear if it refers to TDP, TGP, or just the maximum power of the card. Anyway, it might just be the card’s higher default power (450W+), while the maximum configurable power is the same as on the RTX 4090 Ti (600W). This is the 16 power connector limit.

The faster memory chips say the card will have a maximum bandwidth of 1152GB/s, which is 144GB/s faster than any RTX 4090 currently on the market. It should also be noted that Kopite confirms card number PG136 and PG139. These are the NVIDIA part numbers and Founders Edition card. In other words, the RTX 4090 Ti will have the same card as the RTX 4090.

It’s no secret that NVIDIA has been planning this card for a long time now. One of the first things that leaked before the launch of the RTX 40 was that piece of hardware from the new cooler design, something that was later used for the non-Ti model RTX 4090 instead.

Kopite does not confirm when this map will be released. In fact, there seems to be no reason to launch a new Ti flagship at the moment. On the other hand, NVIDIA should now have a keynote address for GTC in March, which would have been a good place to announce a new TITAN.

Rumored specifications of the NVIDIA GeForce RTX 40 series
VideoCardz.com Titan Ada RTX4090Ti RTX-4090 RTX4080
foundation Ada (TSMC N4) Ada (TSMC N4) Ada (TSMC N4) Ada (TSMC N4)
graphics processing units AD 102 – 450 m 102 M 102 – 300 AD 103 – 300
card number PG137 SKU0 PG139 / 136-SKU310 PG139/136-SKU330 PG139-SKU360
short message 142 142 126 76
CUDA cores 18176 18176 16384 9728
memory 48 GB G6X 24 GB G6X 24 GB G6X 16 GB G6X
memory bus 384 Idols 384 Idols 384 Idols 256 Idols
memory speed 24 Gbps 24 Gbps 21 Gbps 22.4 Gbps
Can ? / 800 W (Max.) ? / 600 W (Max.) 450 watts (default) / 600 watts (maximum) 320W (default) / 355W (maximum)
power connectors 2x 16 pin 1x 16 pin 1x 16 pin 1x 16 pin
launch date for sure for sure October 2022 November 2022

Source: Kopite7kimi

Share:

Facebook
Twitter
Pinterest
LinkedIn

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Recent Articles:

Table of Contents