Shulou(Shulou.com)12/24 Report--

Thank you, Mr. Air, a netizen of CTOnews.com, for your clue delivery! CTOnews.com, November 29 (Xinhua) at Amazon's "AWS re:Invent 2023" event in Las Vegas today, Peter DeSantis, senior vice president of Amazon's computing division, introduced its three cloud services and launched its new ML chip Trainium2, which claims to have an error rate of only 0.1% and twice the energy efficiency of the previous generation, and a Graviton4 chip with 30% better performance than the previous generation.

CTOnews.com noted that Peter DeSantis claimed that Amazon's quantum computing team had made significant progress in quantum error correction, separating bit flip (Bit-flip) from phase reversal (Phase-flip) errors in the form of logical qubits, thus launching the Trainium2 chip.

DeSantis pointed out that the passive error correction method used by the Trainium2 chip has reduced the error rate by a factor of 100, or "0.1 per cent" of 1/1000, while hardware overheating has been reduced by six times compared with expectations. However, although the new chip has made significant progress in quantum error correction, quantum computing is still in a "very early stage".

The picture on the left of ▲ shows the Graviton4 chip and the picture on the right shows the Trainium2 chip. In addition to the Trainium2 chip, Amazon also introduced its three cloud services, namely ElastiCache Serverless, Aurora Limitless Database and Redshift Serverless Next-generation AI-driven scaling and optimizations, all of which have launched preview versions, of which ElastiCache Serverless has entered a stable state for commercial use.

In addition, Amazon also launched the company's fourth custom processor chip, Graviton4, which has a 30% increase in performance, a 50% increase in the number of cores and a 75% increase in memory bandwidth over the previous generation, which can help various workloads of Amazon EC2 cloud services provide the best cost performance and energy efficiency. The processor also encrypts all high-speed physical hardware interfaces to improve security.

According to CTOnews.com earlier, Microsoft launched its own artificial intelligence chip Maia, while Google also has its own TPU tensor computing unit. Now Google, Microsoft, Amazon and other big companies have launched their own customized AI chips, which is also a move made by cloud computing manufacturers to move away from relying on processors made by Intel and AMD.

Related reading: "Microsoft released self-developed AI chips Azure Maia 100 and Cobalt 100 to enhance Azure AI and Copilot services."

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.