[REVIEW] : Nvidia GeForce RTX 40 graphics can consume over 450 watts
After the huge success of the current Nvidia GeForce RTX 30 graphics cards, the North American company is not giving up and is already fully involved in the development of its new range.
And according to the latest information, the upcoming GeForce RTX 40 GPUs can consume over 450 watts of power.
Nvidia GeForce RTX 40 can draw over 450 watts
Nvidia’s new GeForce RTX 40 graphics cards are slated for launch in late 2022. And the brand is really committed to their development as we recently reported that the company has already paid $9 billion to secure the stock. chips for new graphics equipment.
But the new information now released indicates that Nvidia’s next-gen graphics cards can draw over 450 watts of power. So, if this feature is confirmed, the new GPUs may consume more than the current top-end GeForce RTX 3090 Ti model. In turn, the user may even have to invest in a more powerful power supply.
For now, the leading brand in the graphics chip segment has not yet revealed many details about this new line. However, some data already revealed shows that the power consumption of the RTX 40 can vary between 450W and 850W. Popular leaker Greymon claims to have had access to data for the AD102 model, which featured 450W, 650W, and 850W TGPs, though he didn’t say if it’s the same graphics card or if each value pertains a different product.
But the various leakers seem to agree that the Nvidia GeForce RTX 40 will arrive in September this year. The company has previously said that Taiwanese TSMC’s 5nm process promises up to 16% performance improvement over the current 7nm.
NVIDIA will launch the GeForce RTX 40 in September
These graphics cards promise to be energy-hungry: after 600 W TGPs, some are now talking about 800 W, or even 850 W.
After Turing and Ampere, the next generation of NVIDIA graphics cards will use the Ada Lovelace architecture. The GeForce RTX 4000 should benefit from better engraving finesse than the current generation (5 nm instead of 8 nm) and their manufacture would be entrusted to TSMC rather than Samsung. In terms of consumption, however, the TGPs could soar compared to the current range: some mention 800 W or even 850 W for the AD102 GPU.
These values were mentioned by Kopite7kimi and Greymon55 in Tweets posted yesterday. The second also communicates a date. According to him, NVIDIA will launch the first GeForce RTX 4000s next September. If this schedule materializes, two years will have passed between the two generations: the GeForce RTX 3080 had inaugurated the Ampere architecture in September 2020.
450W for the GeForce RTX 3090 Ti
As it stands, we cannot deny or confirm the consumption figures given above. What seems certain is that the RTX 4000s will be more greedy than the RTX 3000s. Several references from this generation will use the 12+4 pin connector, qualified as PCIe 5.0 Ready by the brands. Asus, MSI, Gigabyte and Cooler Master have already drawn power supplies equipped with this type of cable / connector delivering up to 600 watts.
Already, the upcoming GeForce RTX 3090 Ti is suspected to have a 450W TGP. The assumption of GeForce RTX 4080, RTX 4080 Ti and RTX 4090 consuming as much or more (up to 600W in this case) is therefore plausible. On the other hand, 800 or 850 W TGPs for consumer GPUs seem extravagant
While Nvidia’s RTX 3000 generation will have had a complicated history, the manufacturer is preparing the release of its RTX 4000s and announces that it has secured supplies to avoid any shortages.
While the brand is preparing the release of the “SUPER” versions of its RTX 3000s, Nvidia is already thinking above all of securing the supply of the chips necessary for the manufacture of its RTX 4000s.
This is no surprise to anyone: since the release of the RTX 3000 in the midst of a health crisis and shortage of components, it is difficult to get your hands on a next-gen graphics card, and when possible, you have to compose with inflated prices by resellers and wholesalers.
The situation is quite complex for Nvidia, which is singled out in the case, and even more so since the brand in turn increased the price of its Founder Edition.
The RTX 4000s are expected for the start of the 2022 school year and to avoid stock problems, it seems that Nvidia spent close to $9 billion in the last quarter to secure the supply of components. With this money, Nvidia would also have reserved the production load at TSMC at the level of the 5 nm engraving lines, a technology which will be at the heart of the Ada Lovelace architecture.
Nvidia could also spend 3 billion more in the coming quarters to secure production volumes a little more. Logically, Nvidia could also produce many more of its Founder Edition cards. It remains to be seen whether this will be enough to ensure more stocks on the market. Note in passing that no information has circulated regarding a switch to a catalog of 100% LHR models in order to limit the actions of scalpers.
NVIDIA RTX LHR v2 Unlocker: malware is hiding in the software!
A few days ago we saw the arrival of a new utility to unlock 100% of the capabilities of graphics cards NVIDIA RTX 3000 LHR to mine cryptocurrencies. According to the author, the utility somehow recovers the original GPU BIOS. After flashing, the user only has to download and install a special driver. Still according to the author, this simple manipulation completely unlocks NVIDIA’s hash limiter.
Do not download NVIDIA RTX LHR v2 Unlocker tool, it contains malware
Except that in reality, we now learn that this tool contains malware not very nice. We can only advise you to do not download this tool. This malware connects to a remote server and steals user data and uses PC resources for evil purposes.
The Nvidia RTX 4090 would be the most power-hungry graphics card with 850W of TDP
Two famous leakers, Greymon55 and kopite7kimi, announced very bad news for the next generation of Nvidia graphics cards. According to them, GPUs could break energy consumption records since the new RTX 4080, 4080 Ti and 4090 would consume 450W, 650W and 850W respectively because of their new AD102 chip.
By comparison, the GeForce RTX 3090, Nvidia’s most powerful graphics card right now, needs 350W to operate. The new generation would therefore need more than twice as much power, which corroborates previous leaks announcing that Ada Lovelace will offer 2 to 2.5 times more performance than Ampere.
A 1500W POWER SUPPLY FOR THE RTX 4090?
Nvidia should therefore use several PCI Express 5.0 connectors with 16-pin ports, each of which can deliver up to 600W to graphics cards. Ada Lovelace would not be radically different from Ampere, but should use TSMC’s 5nm etching process. This would be 16% faster than 7nm at the same power or 14% more efficient if it delivers the same performance, with a density gain of around 70%.