Nvidia’s next-gen GPUs to be designed with AI and machine learning
Nvidia's AI investments pay off in spades
Share this article...
WePC is reader-supported. When you buy through links on our site, we may earn an affiliate commission. Prices subject to change. Learn more
At GTC, Nvidia’s chief scientist and SVP of research, Bill Daly claimed that the company’s R&D teams will be making use of Artificial Intelligence in addition to machine learning to help bolster their designs and the efficacy of their next-generation GPUs, as reported by WCCFTech. Artificial intelligence and machine learning are one of Nvidia’s strongest points. The company has launched several products in the past for consumers that make great use of their AI and ML modelling such as DLSS, which is by all rights an incredibly disruptive feature in the GPU market. Now, they are looking to apply their AI and ML skills to another place through the addition of Machine Learning and AI aiding in helping Nvidia to presumably optimize the designs of their silicon.
In one highlighted example during the GTC conference, Dally showed off the increases in performance where a design task involving several intricate processes was boosted from taking three hours to just three seconds instead. This is a substantial improvement that highlights just how incredibly powerful AI can be when looking at actually designing its products.
Nvidia shows off AI and ML in their future devices
In addition to this, AI and ML is able to help control other functions such as cell migration, voltage drops and anticipating any errors that might occur during the process. So, you will be able to ensure that your tastes will now be executed in a fraction of the time that they were previously performed. One great example of this is in the voltage drop tools, which increase the accuracy and speed of the tasks that devise where power flows in a system by 94%.
Their circuit designers also use neural networks in order to investigate and locate any issues that would take time out of their development processes. In addition to this is in chip migration, where 92% of the library is now able to be performed without any errors and to quote Nvidia themselves ‘in many cases, we wind up with a better design’.
Will this help Nvidia create future GPUs like Lovelace?
It’s probably a little bit too soon to get Nvidia to apply these ML methods to a product whose designs have already been finalized, meaning that this improvement might actually take a little while to come into the consumer’s hands. This is also not strictly going to mean that you are going to get more performance out of your future GPU, however.
Instead, these improvements seek to serve the development processes of the cards themselves, which in themselves will define just how fast the GPU is able to get. Whatever lies beyond Lovelace and Hopper can expect to enjoy these improvements in Nvidia’s development process, so you will soon be able to get the best of the best when it comes down to getting the best performance out of any GPU that you purchase in the future. Really, this news is going to be interesting to those who have not yet upgraded from their GPUs which are just a few years old, rather than anyone looking for an upgrade immediately, since these improvements again will not be reflected in Nvidia’s newest chip designs, if they are just being revealed now instead of being in one released later this year.
Nvidia’s investment in AI is paying off
Again, Nvidia prides itself on their AI offering, meaning that we could see further improvements and innovations from the Taiwanese giant. We’ve already seen Nvidia develop consumer-facing products such as Nvidia Broadcast, DLSS, DLDSR and more releases, and if their commitment to adding to their GPU offering with more consumer-facing products, we could easily see that there could be further improvements made in this regard with further consumer applications. We don’t know what Nvidia might be cooking up, but if it’s anything on the level of DLSS, we absolutely cannot wait to see them used, while in tandem with chips and silicon that was also designed with the aid of their ever-improving machine learning and AI technologies.
While we might not see any immediate changes on the horizon when it comes to these improvements in design, it’s going to be an incredibly exciting day when we finally see some of these improvements not only bear fruit, but should also put their competitors such as AMD and now Intel on notice.
It’s one of those moves which look to be incredibly long-term, and with GPU stocks beginning to stabilize, you can expect that these graphics cards should be available at launch, just so long as we don’t see another crypto pricing rally that could see every PC part get just a bit more expensive again. The best thing about Nvidia is that they are aware of when they need to be competitive, and know when they are ahead. With the current crop of GPUs coming startlingly close to performance between brands, Nvidia will want to once again get the power and value crowns, all of which are bolstered by their consumer AI product sets.
Why Nvidia needs to keep pricing competitive for consumer GPUs
There is more competition than ever in the GPU ring, meaning that Nvidia will need to leverage every trick available to them in order to get the best performance out of their products, and also grow a good community following, which will inevitably be at the heart of this offering. Therefore, you should take up the mantle of getting the best possible experience all around, not just on pure silicon performance, no matter how important it might actually be. Therefore, one of the best things that they can do is to offer great performance, and their AI offering at a reasonable piece in order to keep up their market momentum, as reflected in the Steam Hardware Survey. This strong showing always comes out as being slightly more expensive than many others on the market, but this is reflective of their extensive offering when it comes to looking at the Nvidia ecosystem as a whole.
Regardless, competitors need to stay on their toes when looking at what Nvidia is able to pull out of the bag with their AI and ML improvements, so keep your eyes peeled on any more news that we might get about new GPUs that we could be seeing coming to a PC near you soon.
