IBM Plans Release of Tesla P100 GPU on Bluemix
IBM plans to release the Tesla P100 as the latest generation of the NVIDIA graphics processing unit (GPU) on its Bluemix bare-metal servers. This will allow businesses to run high-performance workloads like machine learning and artificial intelligence (AI). IBM’s offering will mark an industry first for such a product.
The availability and speed of the Tesla P100 is expected to enable organizations to make sense of large bundles of data and address complex situations more efficiently than they could in the past.
“The Tesla P100 is regarded as the leader of NVIDIA’s GPU pack. It uses the Pascal GPU architecture, which is not only speedier overall than the previous generation of 2012 Kepler-powered processors but also includes new types of GPU instructions to accelerate certain calculations. Software that takes advantage of the Pascal instruction set, like the Torch deep-learning framework, runs even faster,” according to an article in InfoWorld.
And Big Blue is the first major cloud provider to make NVIDIA’s latest technology available worldwide on the cloud, according to an article in VentureBeat.
“IBM has been working in lockstep with NVIDIA for several years and has consistently been first to bring the latest GPU technology to the cloud,” IBM Cloud Infrastructure General Manager John Considine told VentureBeat. “IBM was first to deliver the K80 and M60 GPUs on the cloud in 2015 and 2016, and we are building on that leadership with the release of the P100.”
Bluemix in a Competitive Landscape
Other major cloud vendors offer NVIDIA GPUs as well, but many only offer previous-generation silicon, InfoWorld mentions. For instance, Google Cloud Platform, Microsoft Azure and Amazon Web Services currently use NVIDIA K80s. The Tesla P100 accelerator provides up to 65 percent more machine learning capabilities and 50 times the performance of Tesla K80, according to an article in insideHPC.
With Big Blue’s offering, clients will be able to attach up to two Tesla P100 accelerator cards to individual Bluemix machines. However, IBM only provides support for the P100 on its Bluemix bare-metal servers, not virtual machines like some public cloud competitors offer.
Yet the Tesla P100 via IBM’s connectivity and Bluemix bare-metal servers provides higher throughput than traditional virtualized servers, according to insideHPC.
Overall, the accessibility to P100s is poised to help businesses solidify operationally by allowing them to gain big-data insights and run resource-intensive calculations to yield valuable problem-solving and improved decision-making.
“The latest NVIDIA GPU technology delivered on the IBM Cloud is opening the door for enterprises of all sizes to use cognitive and AI to address complex big-data challenges,” said Considine in an IBM press release. “IBM’s global network of 50 cloud data centers, along with its advanced cognitive and GPU capabilities, is helping to accelerate the pace of client innovation.”
IBM’s offering of Tesla P100 GPU accelerators should be available in May 2017.