At present, I needed to share some essential information on Oracle’s compute infrastructure choices. Listening to our clients about their necessities and experiences, we discovered that the promise of high-performance computing (HPC) and efficiency within the public cloud nonetheless comes with too many trade-offs, caveats, and gaps.

So, now we have updates each to our present capabilities and information associated to our Compute providers roadmap. These bulletins are designed to fulfill the excessive expectations our infrastructure clients, together with for HPC options to ship the most effective of each worlds: absolutely the efficiency they count on from on-premises techniques and the pay-per-use, flexibility, and scalability of cloud, all at a number one value level.

Doubling Down on Future Excessive-Efficiency Computing Investments with Intel

Our current portfolio of HPC capabilities is best-in-class and extra similar to on-premises purpose-built clusters. Companies from different cloud suppliers have underserved this market. During the last yr, we’ve invested on this enviornment to offer clients with the advantages of each on-premises and cloud with not one of the downsides, comparable to overbuilding to peak capability.

And clients have responded. Our recent announcement with Nissan Motors in Japan is a superb instance, the place Nissan moved to Oracle Cloud Infrastructure for its Computational Fluid Dynamics (CFD), crash simulations, and 3D visualization workloads. Altair Engineering announced today that they’re betting on Oracle Cloud Infrastructure as their most well-liked supplier for inner use and for customer-facing SaaS merchandise.

“Our goal is to help customers solve complex problems faster, easier, and smarter than ever before,” mentioned Sam Mahalingam, CTO of Altair. “We’ve standardized on Oracle Cloud Infrastructure, because it delivers the best price-performance, enabling our customers to easily design innovative sustainable products.”

As a part of our HPC platform roadmap, we’re collaborating carefully with Intel, and we’re excited to announce that early subsequent yr, we’re providing our subsequent technology HPC Compute cases primarily based on Intel’s Ice Lake processors. Workloads comparable to crash simulations, CFD, and Digital Design Automation (EDA) workloads are anticipated to ship greater than 30 p.c extra efficiency in comparison with our current X7 HPC technology of cases. Clients also can deploy these cases as naked metallic, get NVMe storage for native check-pointing, a balanced core-memory ratio, and talent to construct clusters of those cases on our distant direct reminiscence entry (RDMA)-enabled cluster community. Lastly, to feed all this computing, you may also construct massive distributed storage clusters working high-performance file techniques, and we have constructed a extremely scalable file system to serve this function. The latest IO500 exhibits that Oracle has the fastest, cloud-based BeeGFS cluster in the world.


Put merely, Oracle Cloud Infrastructure has and continues to have the most effective cloud HPC platform.

NVIDIA A100 Basic Availability

Persevering with our investments in HPC, our current set of NVIDIA GPU choices is best-in-class inside the public cloud house. The truth is, Oracle Cloud Infrastructure’s present NVIDIA GPU infrastructure is extra similar to on-premises techniques, comparable to NVIDIA’s DGX, as an amazing various for patrons that want the best possible efficiency for his or her workloads, comparable to deep studying coaching or {hardware} accelerated visualizations.

We introduced earlier this yr that Oracle is working with NVIDIA on the following technology of GPU cases in Oracle Cloud Infrastructure. Today we’re announcing general availability of these cases beginning September 30 in our US, EMEA, and JAPAC areas, at an trade low on-demand value of $3.05 per GPU per hour. This naked metallic occasion joins our cluster community structure, permitting clients to scale as much as 512 GPUs in a single cluster for large-scale synthetic intelligence coaching or HPC workloads.

A graphic showing a quote from and a picture of Jensen Huang, Founder and CEO of NVIDIA.

These cases present as much as 1.6 Tb per second of bandwidth per bare-metal node that homes eight A100 GPUs, all totally interconnected with NVLINK. These cases additionally offer you over 25 Tb of native NVMe storage and a couple of Tb of RAM for large-scale graph workloads or accelerated databases. Lastly, we additionally enable for leading edge capabilities comparable to GPUDirect via RDMA, not but out there on another cloud supplier. Moreover, you should utilize all the prevailing toolsets, preconfigured Information Science digital machines, and Market pictures, together with assist for NVIDIA GPU Cloud (NGC). We’ve been working with a number of clients in preview over the previous few months, together with DeepZen, IdenTV, and Altair.

“Replicating the human voice with AI is highly dependent on processing power, and Oracle Cloud Infrastructure delivers that with the new NVIDIA A100 GPU, which provided an immediate performance increase of 37% enabling us to scale our business.” Kerem Sozugecer, Co-Founder and CTO and Deep Zen Restricted.

“The amount of streaming video data being created is growing exponentially. To deliver real-time analytics and insights demands the highest level of graphics processing units. Oracle Cloud Infrastructure delivers that with the new NVIDIA A100 GPU where we expect an immediate performance gain of 35%.” Amro Shihadah, Cofounder and COO of IDenTV.

One factor is evident. That is the cloud-based GPU occasion on-premises you’ve been ready for. Discover out extra in NVIDIA and Oracle Cloud Infrastructure NVIDIA GPU Cloud Platform.

Partnership with Rescale

We’re also announcing a partnership with Rescale. They’re one of many main HPC suppliers at the moment. Availability of Rescale on Oracle Cloud Infrastructure makes it even simpler so that you can onboard and get your jobs working in underneath a day. Rescale has greater than 450 purposes already preinstalled on Oracle’s high-performance computing cases, making it simple and supplying you with the flexibility to convey your personal licenses as nicely.

“Rescale takes the complexity out of high-performance computing in the cloud by providing an HPC platform that can be deployed in minutes,’ said Terry Danzer, COO at Rescale. “Oracle Cloud Infrastructure, with its industry first bare metal HPC shapes and low latency RDMA networking, offers the ideal platform for high-performance computing.”

Partnering with Ampere for Oracle’s First Arm Providing

Arm-based compute cases present an amazing various for builders to diversify and benefit from {hardware} innovation. To realize this purpose, we partnered with Ampere to convey an Arm providing to Oracle Cloud. These Arm-based cases present builders with the most effective price-performance, in comparison with another x86 Compute occasion on a per core foundation with an order of magnitude of value financial savings.]

A graphic showing a quote from and picture of Renee Jamesm Chairman and CEO of Ampere.

Early subsequent yr, we’re going to supply clients the flexibility to launch naked metallic or digital machine cases with as much as 160 cores with 3.Three Ghz turbo frequency on extensive number of Linux distros, together with Oracle Linux and Ubuntu. These cases are additionally a part of our versatile Compute household, supplying you with the flexibility to choose and select various ranges of cores or reminiscence primarily based in your workload traits and necessities.

Other than pure Compute cases, you may also use these cases as a part of our At all times Free Tier, permitting you develop and check. Look out for extra info on this providing later within the yr.

Subsequent Technology of Versatile, Elastic Computing

We just lately introduced our E3 cases, which had been the primary technology of our versatile infrastructure. Utilizing the infrastructure, clients can design their cases with a customized variety of CPU cores and reminiscence to swimsuit their workload and software wants, versus getting set into mounted sizes with wasted sources and prices. Constructed on AMD’s Rome technology of CPUs, this service permits you to design customized shapes. For instance, should you needed two cores however 12-GB RAM, it’s now potential to design that customized, actual form to match to the workload. Because the least costly per core Compute providing that we had in our portfolio of cases, it’s an amazing selection for general-purpose workloads.

To get even additional towards our purpose of really responsive, performant, and cost-effective Compute providers, early subsequent yr we’re going to supply clients E4 cases that construct upon this platform with AMD’s subsequent technology of CPUs cases. They permit for higher efficiency on the identical economics, making this service best for general-purpose workloads or customized purposes.

A graphic showing a quote from and picture of Lisa Su, President and CEO of AMD

Get Began At present!

For those who haven’t already, I like to recommend watching Clay’s keynote, together with particular look from Intel CEO Bob Swan, NVIDIA Founder and CEO Jensen Huang, CEO of Ampere Renee James and AMD CEO Dr. Lisa Su.

You additionally get to listen to from clients and companions about how Oracle Cloud Infrastructure helped them remodel, accomplice, and supply the most effective infrastructure out there at the moment.

You may get began at the moment with a few of our newest choices by signing up for our Oracle Cloud Free Tier and get entry to preview of future choices by getting in contact with us immediately.


What is high-performance computing (HPC)?

High Performance Computing (HPC) on Oracle Cloud Infrastructure

Leave a Reply

Your email address will not be published. Required fields are marked *