Advanced Micro Devices (NASDAQ:AMD)
According to a survey published on Friday by an AI software company, Advanced Micro Devices (NASDAQ:AMD) artificial intelligence processors are around 80% as fast as those from Nvidia (NASDAQ:NVDA) Corp, with a potential route to equal their performance. As a result, AMD stock.
In recent months, artificial intelligence (AI) applications like ChatGPT have swept the technology sector, and Nvidia has dominated the market for the powerful CPUs necessary to produce these services. Nvidia’s chip scarcity is attributable to the growing demand for its services, which have propelled the company’s valuation above $1 trillion.
Meanwhile, IT firms hope that AMD stock will emerge as a formidable competitor. As a result, AI firm MosaicML, which was bought by Baidu for $1.3 billion earlier this week, decided to experiment to see how well AMD and Nvidia’s AI processors performed.
The AMD MI250 and the Nvidia A100 were tested by MosaicML since they are one generation below their respective companies’ flagship CPUs but remain popular among consumers.
Thanks to updates to Advanced Micro Devices (NASDAQ:AMD) software at the end of last year and to open-source software funded by Meta Platforms called PyTorch, MosaicML discovered that AMD’s CPU could obtain 80% of the performance of Nvidia’s chip.
MosaicML’s CTO Hanlin Tang said the business expects future software upgrades from AMD to enable its MI250 processor to compete with Nvidia’s A100 in terms of performance.
Tang noted that “for most (machine learning) chip companies out there, the software is the Achilles heel of it,” adding that AMD had not compensated MosaicML to do the study. Regarding software, “AMD stock has done really well.”
Without making any modifications to its core code, MosaicML was able to train a huge language model using the tools it developed, including PyTorch and Advanced Micro Devices (NASDAQ:AMD), according to Tang. “you can already switch to these today,” Tang said of AMD’s processors, “they’re essentially interchangeable” with Nvidia chips.
Instead of paying for access to AI systems from providers like ChatGPT maker OpenAI, firms may use MosaicML’s software to build their own AI systems in-house. The business claims it did the study to show that its clients aren’t limited to using Nvidia chips.
In a statement, Advanced Micro Devices (NASDAQ:AMD) stated, “Mosaic’s results reinforce our strategy of supporting an open and easy-to-implement software ecosystem for AI training and inference on AMD hardware.” The company also said it will continue collaborating with Mosaic to fine-tune its software.
Featured Image: Pexels © Andrey Matveev
