Sign up for our Technology Newsletter 
nVidia NVLINK 2.0 Going In IBM Servers
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
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.
Is IBM Going After HP?
IBM has announced a unified branding for its commerce cloud based enterprise products and services with a presentation at the Smarter Commerce Global Summit in Tampa, Florida.
Hot on the heels of HP, which unified its cloud offerings under the Helion brand last week, IBM Experienceone is designed to allow companies to improve engagement with their customers by leveraging big data through the cloud.
Deployment comes from a unified offer of consulting services, software and infrastructure from IBM subsidary Softlayer, which can be used to gather data, mine analytics and improve customer commerce via a mixture of traditional and cloud services.
IBM has already committed 1,000 new employees for its IBM Interactive Experience who will staff 10 “IBM Interactive Experience Labs” that are being set up to help customers understand the rules of engagement and hopefully increase their level of customer engagement.
IBM GM of Industry Cloud Solution Craig Hayman said, “IBM Experienceone provides a secure and simplified portfolio – including innovation from more than 1,200 partners – to help clients design and deliver more valuable customer engagements. With cloud, on premise and hybrid options, IBM Experienceone quickly scales to engage every customer in the moment while protecting their privacy.”
The IBM Experienceone brand is a coming together of many acquisitions that IBM has made in the field over recent years, including Sterling Commerce, Tealeaf, Coremetrics, Unica, Demandtec, Xtify and Silverpop. The only obvious omission from the top to tail offer is a specific CRM database, however IBM Experienceone is compatible with most of the leading solutions, including those of its arch rivals. This leads to the question, could a CRM be next on the company’s shopping list?
As well as on desktop and server equipment, Experienceone analytics will also be available through apps for iOS and Android.
IBM To Become Cloud Broker
IBM is in the throes of developing software that will allow organizations to use multiple cloud storage services interchangeably, reducing dependence on any single cloud vendor and ensuring that data remains available even during service outages.
Although the software, called InterCloud Storage (ICStore), is still in development, IBM is inviting its customers to test it. Over time, the company will fold the software into its enterprise storage portfolio, where it can back up data to the cloud. The current test iteration requires an IBM Storewize storage system to operate.
ICStore was developed in response to customer inquiries, said Thomas Weigold, who leads the IBM storage systems research team in IBM’s Zurich, Switzerland, research facility, where the software was created. Customers are interested in cloud storage services but are worried about trusting data with third party providers, both in terms of security and the reliability of the service, he said.
The software provides a single interface that administrators can use to spread data across multiple cloud vendors. Administrators can specify which cloud providers to use through a point-and-click interface. Both file and block storage is supported, though not object storage. The software contains mechanisms for encrypting data so that it remains secure as it crosses the network and resides on the external storage services.
A number of software vendors offer similar cloud storage broker capabilities, all in various stages of completion, notably Red Hat’s DeltaCloud and Hewlett Packard’s Public Cloud.
ICStore is more “flexible,” than other approaches, said Alessandro Sorniotti, an IBM security and cloud system researcher who also worked on the project. “We give customers the ability to select what goes where, depending on the sensitivity and relevance of data,” he said. Customers can store one copy of their data on one provider and a backup copy on another provider.
ICStore supports a number of cloud storage providers, including IBM’s SoftLayer, Amazon S3 (Simple Storage Service), Rackspace, Microsoft Windows Azure and private instances of the OpenStack Swift storage service. More storage providers will be added as the software goes into production mode.
“Say, you are using SoftLayer and Amazon, and if Amazon suffers an outage, then the backup cloud provider kicks in and allows you to retrieve data,” from SoftLayer, Sorniotti said.
ICStore will also allow multiple copies of the software to work together within an enterprise, using a set of IBM patent-pending algorithms developed for data sharing. This ensures that the organization will not run into any upper limits on how much data can be stored.
IBM has about 1,400 patents that relate to cloud computing, according to the company.
IBM Buys SoftLayer
IBM has signed an agreement to purchase SoftLayer Technologies, as it looks to accelerate the build-out of its public cloud infrastructure. The company is also forming a services division to back up the push.
The financial details of the deal were not announced, but SoftLayer is the world’s largest privately held cloud computing infrastructure provider, according to IBM.
IBM already has an offering that includes private, public and hybrid cloud platforms. The acquisition of SoftLayer will give it a more complete in-house offering, as enterprises look to keep some applications in the data center, while others are moved to public clouds.
SoftLayer has about 21,000 customers and an infrastructure that includes 13 data centers in the U.S., Asia and Europe, according to IBM. SoftLayer allows enterprises to buy compute power on either dedicated or shared servers.
Following the close of the acquisition of SoftLayer, which is expected in the third quarter, a new division will combine its services with IBM’s SmartCloud. IBM expects to reach $7 billion annually in cloud revenue by the end of 2015, it said.
Success is far from certain: The public cloud market is becoming increasingly competitive as dedicated cloud providers, telecom operators and IT vendors such as Microsoft and Hewlett-Packard all want a piece. The growing competition should be a good thing for customers if it drives down prices. For example, Microsoft has already committed to matching Amazon Web Services prices for commodity services such as computing, storage and bandwidth.
Not all hardware vendors feel it’s necessary to have their own public cloud. Last month, Dell changed strategy and said it would work with partners including Joyent, instead of having its own cloud.