Availability of Azure OpenAI service will be expanded as Microsoft announced yesterday

Microsoft yesterday announced the expanded availability of its Azure OpenAI service. With this expansion, customers in Australia East, Canada East, US East 2, Japan East, and UK South will be able to access the popular OpenAI models GPT-4 and GPT-35-Turbo. Prior to this expansion, the Azure OpenAI service was available in US East, France Central, US South Central, and Western Europe. During its recent earnings call, Microsoft announced that the Azure OpenAI service now serves more than 11,000 customers, attracting an average of 100 new customers per day.

Microsoft also announced the general availability of the Azure ND H100 v5 virtual machine (VM) series, which features the latest NVIDIA H100 Tensor Core GPU and NVIDIA Quantum-2 InfiniBand networking. This new VM series is designed for AI workloads and is now available in the US East and US South Central Azure regions. Even if a VM is officially released, customers must register their interest in accessing the new VM.

ND H100 v5 virtual machines include the following features:

• These virtual machines are equipped with eight NVIDIA H100 Tensor Core GPUs, promising to implement AI models faster than previous generations. performance.
• 4th generation Intel Xeon Scalable processors are the foundation of these virtual machines, ensuring optimal processing speeds.
• NVIDIA Quantum-2 ConnectX-7 InfiniBand 400Gb/s per GPU and 3.2 Tb/s of cross-node bandwidth per virtual machine ensures seamless performance across GPUs, matching the capabilities of the world’s top supercomputers .
• With PCIe Gen5 providing 64GB/s of bandwidth per GPU, Azure achieves significant performance advantages between CPUs and GPUs.
• DDR5 memory is at the heart of these virtual machines, delivering higher data transfer speeds and efficiency, making it ideal for workloads with larger data sets.
• Matrix multiplication operations are accelerated by a factor of six when using the new 8-bit FP16 floating-point data type compared to FP8 in previous generations.
• Achieved a twofold speedup in large language models such as BLOOM 175B end-to-end model inference, demonstrating their potential to further optimize AI applications.

The above is the detailed content of Availability of Azure OpenAI service will be expanded as Microsoft announced yesterday. For more information, please follow other related articles on the PHP Chinese website!