Syber Group
Toll Free : 855-568-TSTG(8784)
Subscribe To : Envelop Twitter Facebook Feed linkedin

nVidia NVLINK 2.0 Going In IBM Servers

August 31, 2016 by  
Filed under Computing

Comments Off on nVidia NVLINK 2.0 Going In IBM Servers

On Monday, PCWorld reported that the first servers expected to use Nvidia’s second-generation NVLINK 2.0 technology will be arriving sometime next year using IBM’s upcoming Power9 chip family.

IBM launched its Power8 lineup of superscalar symmetric multiprocessors back in August 2013 at the Hot Chips conference, and the first systems became available in August 2014. The announcement was significant because it signaled the beginning of a continuing partnership between IBM and Nvidia to develop GPU-accelerated IBM server systems, beginning with the Tesla K40 GPU.

The result was an HPC “tag-team” where IBM’s Power8 architecture, a 12-core chip with 96MB of embedded memory, would eventually go on to power Nvidia’s next-generation Pascal architecture which debuted in April 2016 at the company’s GPU Technology Conference.

NVLINK, first announced in March 2014, uses a proprietary High-Speed Signaling interconnect (NVHS) developed by Nvidia. The company says NVHS transmits data over a differential pair running at up to 20Gbps, so eight of these differential 20Gbps connections will form a 160Gbps “Sub-Link” that sends data in one direction. Two sub-links—one for each direction—will form a 320Gbps, or 40GB/s bi-directional “Link” that connects processors together in a mesh framework (GPU-to-GPU or GPU-to-CPU).

NVLINK lanes upgrade from 20Gbps to 25Gbps

IBM is projecting its Power9 servers to be available beginning in the middle of 2017, with PCWorld reporting that the new processor lineup will include support for NVLINK 2.0 technology. Each NVLINK lane will communicate at 25Gbps, up from 20Gbps in the first iteration. With eight differential lanes, this translates to a 400Gbps (50GB/s) bi-directional link between CPUs and GPUs, or about 25 percent more performance if the information is correct.

NVLINK 2.0 capable servers arriving next year

Meanwhile, Nvidia has yet to release any NVLINK 2.0-capable GPUs, but a company presentation slide in Korean language suggests that the technology will first appear in Volta GPUs which are also scheduled for release sometime next year. We were originally under the impression that the new GPU architecture would release in 2018, as per Nvidia’s roadmap. But a source hinted last month that Volta would be getting 16nm FinFET treatment and may show up in roughly the same timeframe as AMD’s HBM 2.0-powered Vega sometime in 2017. After all, it is easier for Nvidia to launch sooner if the new architecture is built on the same node as the Pascal lineup.

Still ahead of PCI-Express 4.0

Nvidia claims that PCI-Express 3.0 (32GB/s with x16 bandwidth) significantly limits a GPU’s ability to access a CPU’s memory system and is about “four to five times slower” than its proprietary standard. Even PCI-Express 4.0, releasing later in 2017, is limited to 64GB/s on a slot with x16 bandwidth.

To put this in perspective, Nvidia’s Tesla P100 Accelerator uses four 40GB/s NVLINK ports to connect clusters of GPUs and CPUs, for a total of 160GB/s of bandwidth.

With a generational NVLINK upgrade from 40GB/s to 50GB/s bi-directional links, the company could release a future Volta-based GPU with four 50GB/s NVLINK ports totaling of 200GB/s of bandwidth, well above and beyond the specifications of the new PCI-Express standard.

Courtesy-Fud

IBM Goes Bare Metal

March 18, 2015 by  
Filed under Computing

Comments Off on IBM Goes Bare Metal

IBM has announced the availability of OpenPower servers as part of the firm’s SoftLayer bare metal cloud offering.

OpenPower, a collaborative foundation run by IBM in conjunction with Google and Nvidia, offers a more open approach to IBM’s Power architecture, and a more liberal licence for the code, in return for shared wisdom from member organisations.

Working in conjunction with Tyan and Mellanox Technologies, both partners in the foundation, the bare metal servers are designed to help organisations easily and quickly extend infrastructure in a customized manner.

“The new OpenPower-based bare metal servers make it easy for users to take advantage of one of the industry’s most powerful and open server architectures,” said Sonny Fulkerson, CIO at SoftLayer.

“The offering allows SoftLayer to deliver a higher level of performance, predictability and dependability not always possible in virtualised cloud environments.”

Initially, servers will run Linux applications and will be based on the IBM Power8 architecture in the same mold as IBM Power system servers.

This will later expand to the Power ecosystem and then to independent software vendors that support Linux on Power application development, and are migrating applications from x86 to the Power architecture.

OpenPower servers are based on open source technology that extends right down to the silicon level, and can allow highly customised servers ranging from physical to cloud, or even hybrid.

Power systems are already installed in SoftLayer’s Dallas data centre, and there are plans to expand to data centres throughout the world. The system was first rolled out in 2014 as part of the Watson portfolio.

Prices will be announced when general availability arrives in the second quarter.

Source