NVIDIA is currently celebrating its GTC, but it has focused more on markets outside of the PC world where Jensen Huang’s company has grown in importance since moving from a graphics hardware designer to that of artificial intelligence. But what about future graphics cards like the RTX 30 SUPER or the next generation with the RTX 40 based on the Ada Lovelace architecture?
The RTX 4090 will consume 600W
A few days ago, we explained to you how the new 12-pin power supply port that will be released in PCI Express 5.0 will allow the cards to be supplied with 600 W of power, a connector which is none other than the same connector that NVIDIA already had. . included in some models of their RTX 30 and so we shouldn’t be surprised that it is being used at its best in its next generation of graphics cards for gaming.
What do we know about the RTX 40? Well, little really, the most reliable sources say that they will use the TSMC 5nm node to make the chip and that it will have a total of 144 SM, an impressive number that indicates a mammoth chip. Of course, such a level of MS also means having an even larger communication infrastructure than the current one. The results ? The fact that we are going to go from the 350 W of the RTX 3090 to the 550 to 600 W of consumption of the RTX 4090.
RTX 4090 release date
The chip the RTX 4090 will use as a GPU is the AD102, a beast that will surely touch the edge of the reticle since it is a large monolithic chip. Under the 7nm node we have already seen AMD place 128 SM with its Intinct MI200 on a single chip, so it is not impossible that Nvidia succeeds in placing 144 units under a node which, being conservative, has a density between 1, 6 and 2 times larger.
But what interests us is to know its launch date and it seems that NVIDIA has become quite classic in this, since we will see the first models appear in the third quarter of 2022, only two years after the presentation of the RTX 30. In other words, we could see them in the market from July 2022.
This means that the launch distance between the RTX 30 SUPER and the early models of the RTX 40 will be very narrow. Our prediction? Rare and expensive units and Jensen Huang’s company will charge you extra for that extra level of power.
RTX 4090 with lower performance per watt?
NVIDIA has been increasing the performance per watt of its graphics cards for several generations, which means that for the same power consumption, the new architectures offer higher performance. Unfortunately, if it was confirmed that NVIDIA with the RTX 40 would have decided to sacrifice the efficiency path in its architectures to brute force, but that would make sense and it has to do with mining, a less efficient architecture in this aspect would make it less attractive to minors.
At the moment it is still too early to talk about the Lovelace architecture and that the RTX 4090 can consume 600 W of power, since what is expected in order to reduce the energy consumption of internal logistics that NVIDIA is developing a new organization. different from that of GPC-TPC-SM that we have seen for over ten years.
Hopper planned for 2022
Let us not forget that for the moment the multichip GPU for HPC under the code name Hopper is still a promise of the future. Meanwhile, AMD with its Instinct MI200 and Intel with its Ponte Vecchio presented prototypes and one of them even a fully functional version. Also, let’s not forget that one of Bill Dally’s biggest obsessions at NVIDIA has always been the power consumption of the operations that take place there, especially those related to data logistics.
The Hopper architecture is supposed to appear next year, but given the delay between the V100 and the A100, a possible H100 could take at least a little longer to appear. In any case, the use of monolithic chips has a scalability limit and sooner or later NVIDIA will have to follow in the footsteps of its competitors by offering a multi-chip GPU.
The latest information indicates that Hopper could appear at the same time as Lovelace, concurrent launches are not normal in this industry, so we could see it at some point in the first half of 2022 or already in the fourth quarter. NVIDIA typically runs two GTCs per year, so Hopper can be in one of them.
Table of Contents
