#gaudi3

Introducing Intel Tiber AI Cloud

Today Intel excited to announce the rebranding of Intel Tiber AI Cloud, the company’s rapidly expanding cloud offering. It is aimed to providing large-scale access to Intel software and computation for AI installations, support for the developer community, and benchmark testing. It is a component of the Intel Tiber Cloud Services portfolio.

What is Intel Tiber AI Cloud?

Intel’s new cloud service, the Intel Tiber AI Cloud, is designed to accommodate massive AI workloads using its Gaudi 3 CPUs. Rather than going up against AWS or other cloud giants directly, this platform is meant to serve business clients with an emphasis on AI computing. Tiber AI Cloud is particularly designed for complicated AI activities and offers scalability and customization for enterprises by using Intel’s cutting-edge AI processors.

Based on the foundation of Intel Tiber Developer Cloud, Intel Tiber AI Cloud represents the rising focus on large-scale AI implementations with more paying clients and AI partners, all the while expanding the developer community and open-source collaborations.

New Developments Of Intel Tiber AI Cloud

It have added a number of new features to Intel Tiber AI Cloud in the last several months:

New computing options

Nous began building up Intel’s most recent AI accelerator platform, Intel Tiber AI Cloud, in conjunction with last month’s launch of Intel Gaudi 3. We anticipate broad availability later this year. Furthermore, pre-release Dev Kits for the AI PC using the new Intel Core Ultra processors and compute instances with the new 128-core Intel Xeon 6 CPU are now accessible for tech assessment via the Preview catalog in the Intel Tiber AI Cloud.

Open source model and framework integrations

With the most recent PyTorch 2.4 release, Intel revealed support for Intel GPUs. Developers may now test the most recent Pytorch versions supported by Intel on laptops via the Learning catalog.

The most recent version of Meta’s Llama mode collection, Llama 3.2, has been deployed in Intel Tiber AI Cloud and has undergone successful testing with Intel Xeon processors, Intel Gaudi AI accelerator, and AI PCs equipped with Intel Core Ultra CPUs.

Increased learning resources

New notebooks introduced to the Learning library allow you to use Gaudi AI accelerators. AI engineers and data scientists may use these notebooks to run interactive shell commands on Gaudi accelerators. Furthermore, the Intel cloud team collaborated with Prediction Guard and DeepLearning.ai to provide a brand-new, well-liked LLM course powered by Gaudi 2 computing.

Partnerships

It revealed Intel’s collaboration with Seekr, a rapidly expanding AI startup, at the Intel Vision 2024 conference. To assist businesses overcome the difficulties of bias and inaccuracy in their AI applications, Seekr is now training foundation models and LLMs on one of the latest super-compute Gaudi clusters in Intel’s AI Cloud, which has over 1,024 accelerators. Its dependable artificial intelligence platform, SeekrFlow, is now accessible via Intel Tiber AI Cloud‘s Software Catalog.

How it Works

Let’s take a closer look at the components that make up Intel’s AI cloud offering: systems, software, silicon, and services.

Silicon

For the newest Intel CPUs, GPUs, and AI accelerators, Intel Tiber AI Cloud is first in line and Customer Zero, offering direct support lines into Intel engineers. Because of this, Intel’s AI cloud is able to provide early access to inexpensive, perhaps non-generally accessible access to the newest Intel hardware and software.

Systems

The size and performance needs of AI workloads are growing, which causes the demands on computational infrastructure to change quickly.

To implement AI workloads, Intel Tiber AI Cloud provides a range of computing solutions, including as virtual machines, dedicated systems, and containers. Furthermore, it have recently implemented high-speed Ethernet networking, compact and super-compute size Intel Gaudi clusters for training foundation models.

Program

Intel actively participates in open source initiatives like Pytorch and UXL Foundation and supports the expanding open software ecosystem for artificial intelligence. Users may access the newest AI models and frameworks in Intel Tiber AI Cloud with to dedicated Intel cloud computing that is assigned to help open source projects.

Services

The way in which developers and enterprise customers use computing AI services varies according on use cases, technical proficiency, and implementation specifications. To satisfy these various access needs, Intel Tiber AI Cloud provides server CLI access, curated interactive access via notebooks, and serverless API access.

Why It Is Important

One of the main components of Intel’s plan to level the playing field and eliminate obstacles to AI adoption for developers, enterprises, and startups is the Intel Tiber AI Cloud:

Cost performance: For the best AI implementations, it provides early access to the newest Intel computing at a competitive cost performance.

Scale: From virtual machines (VMs) and single nodes to massive super-computing Gaudi clusters, it provides a scale-out computing platform for expanding AI deployments.

Open source: It ensures mobility and prevents vendor lock-in by providing access to open source models, frameworks, and accelerator software supported by Intel.