RTX 40 Presenting: Same Power, Better Graphics

GPUs in GeForce RTX 40 will use the same power as non-SUPER cards

It is reported that the power consumption of NVIDIA’s future GeForce RTX 40 SUPER family of GPUs would be comparable to that of the Non-SUPER family.

While providing faster performance, the NVIDIA GeForce RTX 40 SUPER and RTX 40 Non-SUPER GPUs will have very similar power consumption figures.

In response to a previous rumor about the details of NVIDIA’s planned GeForce RTX 40 “SUPER” refresh GPU, Kopite7kimi has now disclosed that the new chips would have power consumption numbers that are strikingly comparable to those of the current Non-SUPER cards.

Again, Kopite7kimi himself said that the specs he possesses are “several weeks” old, so they might change. It is well known that NVIDIA tests a number of configurations before deciding on one.Now that we’ve covered that, let’s discuss GPUs.

SUPER Specifications for NVIDIA GeForce RTX 4080

First up is the NVIDIA GeForce RTX 4080 SUPER, which is said to be equipped with the entire 10,240 cores, 320 TMUs, 112 ROPs, and 64 MB of L2 cache of the AD103-400 GPU. Utilizing a 256-bit bus interface, the chip is expected to include 16 GB of GDDR6X memory, the same amount as the RTX 4080 Non-SUPER. That implies that NVIDIA offers 16 GB (AD103 / 256-bit) and 20 GB (AD102 / 320-bit) SKUs, although it’s unclear whether any of these will be sold in stores. With a TBP of 320W for the NVIDIA GeForce RTX 4080 Non-SUPER, we may anticipate comparable numbers for the 4080 SUPER.

SUPER Specifications for NVIDIA GeForce RTX 4070 Ti

The NVIDIA GeForce RTX 4070 Ti SUPER is the next product on the list. Kopite7kimi thinks the name is strange, and to be honest, it sort of is. Regarding the specs, however, the card is said to include 48 MB of L2 cache and AD103-275 / AD102-175 GPU SKUs with the same 8,448 cores in both versions. The TBP of the GeForce RTX 4070 Ti SUPER ought to be around the 285W threshold of the 4070 Ti.

SUPER Specifications for NVIDIA GeForce RTX 4070

The NVIDIA GeForce RTX 4070 SUPER is the last card on the list. It is said to be available in AD104-350 or AD103-175 SKU configurations, with 7,168 cores shared by both versions. L2 cache on the GPU is anticipated to be 48 MB. Although we may anticipate NVIDIA to increase memory speeds using superior Micron DRAM, which can lessen the strain on the lower 192-bit bus interface, it seems that this card will continue with the 12 GB GDDR6X memory. The TBP of the RTX 4070 SUPER will be 20W more than that of the current 4070, which is 220W.

In conclusion, the core count increase indicated by the rumor is as follows:

• Comparison between RTX 4080 SUPER with RTX 4080 Core Count: +5% augmentation (10240 versus 9728)
• RTX 4070 Ti Core Count = +10% compared to RTX 4070 Ti SUPER Growth (8448 vs 7680)
• Comparison between RTX 4070 SUPER with RTX 4070 Core Count: +22% Rise (7168 as opposed to 5888)

From these, it seems that the NVIDIA GeForce RTX 4070 SUPER would be the most significant improvement; this might be because the AMD RX 7800 XT is perhaps the genuine competitor in this market.

It is recommended that the 4070 Ti SUPER get an increase of 10% in cores, while the RTX 4080 SUPER may have the lowest update, with a 5% increase in core count. This implies that there is still room to add an RTX 4080 Ti to the portfolio.

Better performance at the same power would suggest that the SUPER GPUs are either optimizing or refining the 4N manufacturing node, or that improving the efficiency of the current SKUs would do the trick.

The most energy-efficient gaming chips are now NVIDIA’s Ada Lovelace GPUs, and it seems that the SUPER series will further increase this advantage. A hypothetical debut in 2024 is still a good several months away, so it’s best to take everything with a grain of salt since a lot may happen in that time.

NVIDIA GeForce RTX 40 Desktop GPU Family: