DeepSeek-R1 It is a generative artificial intelligence as powerful as GPT-4o of Open AI, open source, free, that Consume dozens of times less resourcesand that can work at home. It is not strange Openai and Nvidia are in shock, and that Microsoft has already integrated it in Windows 11. Now we also know that pay more on AMD Radeon cardsthat in the powerful RTX 4090 of Nvidia.
The IA Chinese DeepSeek The monopoly of Openai and Nvidia has broken in just a few days At an industrial level, but it seems that it will also do it at domestic level. AMD herself has published some benchmark where Your cards Radeon They offer up to 34% more yield than NVIDIA RTX 4090that until yesterday, that the RTX 5090 was put on sale, it was its most powerful card.
Deepseek-R1 can work at a home on a PC With a moderately powerful graphics card, which is a great advance with respect to co -ilot or GPT-4 models that force you to use the Microsoft or OpenAi cloud.
Depseek works best on Radeon cards
For AMD, Deepseek has been a blessing. Until now I could not compete with Nvidia in AI, but now everything changes. So it has not been cut when promoting Depseek.
David Mcafee The vice president of the department of Ryzen and Radeon in AMD, has hung several graphics where he shows the superior performance of the Radeon cards against the most powerful of NVIDIA, when Deepseek-R1 is used.
Here you can see the results of the benchmarks, Comparing an AMD Radeon 7900 XTX with an RTX 4080 and an RTX 4090:
As can be seen in the graphs, RADEON 7900 XTX is between 22 and 34% faster than RTX 4080according to the model and the parameters of Deepseek-R1 that are used.
It is also between 2 and 13% faster than RTX 4090except in the highest model, with 32,000 million parameters, where the Nvidia card expires by 4% difference.
They are very good results for AMD, which can now boast over Nvidia in AI, at least with Deepseek. On this website, explain how to use China at home with Ryzen Radeon and processed cards endowed with AI chips.
DeepSeek Not only requires much less process power, but It works best on AMD graphics cards than in Nvidia. The scenario has changed completely.
Know How we work in NoticiasVE.
Tags: Nvidia, artificial intelligence, amd, viral