Microsoft is working on a new Artificial Intelligence (AI) infrastructure service codenamed ‘Singularity’ that will offer data scientists and AI researchers. Use a method to scale, build, iterate, and experiment on their models on a Microsoft Azure Cloud service made particularly for AI.
A group of researchers, mostly Indians, have disclosed a paper that offers technical details about the project ‘Singularity.’
“Lowering costs by driving high utilization across deep learning workloads is a crucial lever for cloud providers. We present Singularity, Microsoft’s globally distributed scheduling service for highly-efficient and reliable execution of deep learning training and interference workloads,” researchers wrote.
‘Singularity’ is a completely managed, globally distributed infrastructure service for AI workloads at Microsoft, with support for various diverse hardware accelerators. It has been designed from the ground up to scale across a global fleet of hundreds of thousands of GPUs and other AI accelerators.
As per the researchers, ‘Singularity’ obtains a considerable development in scheduling deep learning workloads, changing niche features like elasticity into the mainstream, always-on attribute that the scheduler can depend on for executing stringent service level agreements (SLAs).
In 2020, Microsoft announced a new supercomputer collaborating with AI, making new infrastructure available in Azure to train extensive AI models. The supercomputer developed for OpenAI is a single system with more than 285,000 CPU cores, 10,000 GPUs, and 400 gigabits per second network connectivity for each GPU server.
Singularity to Time and Cost Required to Train the ML Models
As a result, ‘Singularity’ could eliminate the need to restart the DNN (Deep Neural Networks) training processes from scratch caused due to sudden system failure. Microsoft claims that the process could be resumed from its position when it was preempted. The new service, Singularity, is a significant achievement, and it should assist in lessening the time and effort demanded to train the machine learning models.
Since the past few years, Microsoft has increased its investments in AI and Azure. In 2019, the company disclosed USD 1 Bn investment in OpenAI and expressed its desire to become its partner to commercialize new AI technologies.