Yet another change of specifications for RTX 4000
It has been a few days since we reported the latest alleged technical specifications of the next series of GeForce RTX 40 graphics cards, based on the Ada Lovelace architecture. Recall that, according to the latest rumors, NVIDIA should launch the GeForce RTX 4090 first, and then distribute the GeForce RTX 4080 and GeForce RTX 4070.According to what is reported by the well-known insider kopite7kimi, the GeForce RTX 4090 would be equipped with AD102-300-A1, NVIDIA's most powerful chip whose full version will only be available in the “Ti” variant. The smooth version will be equipped with 16,384 CUDA cores, with a base frequency of 2,235 MHz and a boost of 2,520 MHz (up to a maximum of over 2,750MHz). The card is expected to make use of 24GB of GDDR6X VRAM with 21 Gb / s bandwidth and 384-bit bus interface. The TBP is estimated to be 450W, while a 600W 16-pin power connector should allow it to exceed 500W of power in versions made by third-party partners.
Continue to update.
RTX 4090, AD102-300-A1, 16384FP32, 384bit 21Gbps 24G GDDR6X, 450W, base 2235 boost 2520 actual max> 2750;
RTX 4080, AD103-300-A1, 10240FP32, 256bit 21Gbps 16G GDDR6X, 420W,
RTX 4070, AD104-275-Kx (x is a number) -A1, 7168FP32, 160bit 18Gbps 10G GDDR6, 300W.
- kopite7kimi (@ kopite7kimi) July 4, 2022
The GeForce RTX 4080 will instead be equipped with an AD103-300-A1 graphics processor , with 80 SMs enabled including 10,240 CUDA cores and 224 ROPs, along with a 64MB L2 cache. The memory continues to use modules from GDDR6X, but drops to 16 GB with a speed of 18 Gb / s and the bus at 256-bit. The card is expected to have a TBP of 420W, which is very close to the top-of-the-range model, which will likely allow you to push more frequencies.| ); }
Currently, the Californian company is expected to present the RTX 40 series in September and October. After these first three models, the GeForce RTX 4060 is expected to debut next year at CES 2023.
Nvidia RTX 4000 GPUs might draw more power in gaming laptops and PCs
The Nvidia RTX 4000 rumour mill has been swirling for months now, with countless reports that upcoming GPUs likely won’t so much sip power as they will gorge on it. This increased consumption apparently isn’t limited to gaming PCs either, with team green’s gaming laptop graphics cards seemingly having a larger appetite for watts too.
The flagship GPU of the Nvidia RTX 4000 series, the AD102 which will likely be used in the RTX 4090, will be capable of drawing up to a staggering 800W according to hardware leaker kopite7kimi. This is almost double the 450W power limit found in the company’s current best graphics card, the GeForce RTX 3090 Ti.
A truth. The power limits:AD102, 800W;AD103 (DT), 450W, AD103 (Mobile), 175W;AD104 (DT), 400W, AD104 (Mobile), 175W;AD106 (DT), 260W, AD106 (Mobile), 140W.But I don't think we need to use the full power cap.
— kopite7kimi (@kopite7kimi) June 18, 2022
Moving down the stack, this trend continues with AD103, AD104, and AD106 GPUs, with a floor of 260W and ceiling of 450W. All of this means that picking up an RTX 4080, RTX 4070, or RTX 4060 could result in you needing to pick up one of the best power supplies on the market.
Power limits don’t seem to be so dire on the mobile side of things, with most RTX 4000 gaming laptop graphics cards drawing the same amount of power as their current RTX 3000 equivalents. However, any portable PC packing an AD104 pixel pusher will need to find an additional 10W versus a GA103 GPU.
How much these increased power limits will help or hinder Nvidia’s ability to compete with AMD RDNA 3 graphics cards remains to be seen, but it looks like team green will be the first to market, despite reports of a delay. Although, it’ll be longer still before we’re able to buy RTX 4000 laptops.