Microsoft investment in ChatGPT does not just include money sunk into its maker, OpenAI, but a huge hardware financial investment in data centers also which reveals that in the meantime, AI options are simply for the extremely leading tier companies.The partnership between Microsoft and OpenAI dates back to 2019, when Microsoft invested $1 billion in the AI developer. It upped the ante in January with the financial investment of an additional $10 billion.But ChatGPT has to run on something, which is
Azure hardware in Microsoft information centers. How much has not been disclosed, however according to a report by Bloomberg, Microsoft had actually already invested”a number of hundred million dollars”in hardware utilized to train ChatGPT.In a pair of article, Microsoft detailed what went into constructing the AI infrastructure to run ChatGPT as part of the Bing service. It currently used virtual machines for AI processing developed on Nvidia’s A100 GPU, called ND A100 v4.
Now it is introducing the ND H100 v5 VM based on newer hardware and offering VM sizes ranging from 8 to countless NVIDIA H100 GPUs.In his article, Matt Vegas, principal item supervisor of Azure HPC+AI, composed consumers will see significantly quicker performance for AI designs over the ND A100 v4 VMs. The new VMs are powered by Nvidia H100 Tensor Core GPUs (“Hopper “generation) interconnected by means of
next gen NVSwitch and NVLink 4.0, Nvidia’s 400 Gb/s Quantum-2 CX7 InfiniBand networking, 4th Gen Intel Xeon Scalable processors(“Sapphire Rapids”)with PCIe Gen5 interconnects and DDR5 memory.Just just how much hardware he did not state, but he did state that Microsoft is delivering numerous exaFLOPs of supercomputing power to Azure clients. There is only one exaFLOP supercomputer that we understand of , as reported by the most current TOP500 semiannual list of the world’s fastest: Frontier at the Oak Ridge National Labs. However that’s the important things about the TOP500; not everybody reports their supercomputers, so there might be other systems out there just as powerful as Frontier, however we simply do not learn about them.
In a separate post, Microsoft spoke about how the company began dealing with OpenAI to help develop the supercomputers that are required for ChatGPT’s big language model( and for Microsoft’s own Bing Chat. That meant connecting countless GPUs together in a new way that even Nvidia had not thought of, according to Nidhi Chappell, Microsoft head of product for Azure high-performance computing and AI.”This is not something that you simply purchase an entire bunch of GPUs, hook them together, and they’ll begin collaborating. There is a great deal of system-level optimization to get the best efficiency, which includes a great deal of experience over many generations,”Chappell said.
