Back to top

Image: Bigstock

Oracle (ORCL) Bets on NVIDIA, Ampere for OCI Compute Instances

Read MoreHide Full Article

Oracle (ORCL - Free Report) has announced the upcoming availability for Oracle Cloud Infrastructure (“OCI”) Compute instances that will incorporate advanced hardware components, including next-generation NVIDIA (NVDA - Free Report) GPUs and Ampere-owned AmpereOne CPUs to enhance performances for various workloads.

These instances are designed to cater to a wide range of computing needs, from artificial intelligence (AI) model training and inference to cloud-native applications and video transcoding.

For the second quarter of fiscal 2024, total cloud growth, excluding Cerner, is expected between 27% and 29% at cc. In U.S. dollars, cloud growth is expected between 29% and 31%. Shares of Oracle have gained 38% year to date compared with the Zacks Computer and Technology sector’s rise of 37.3%.

Several organizations, such as Cohere, Databricks and Uber, have already expressed their interest in leveraging these upcoming OCI Compute instances to power the AI and cloud workloads, citing their reliability, performance and scalability.

NVIDIA H100 Tensor Core GPU Instances on OCI

These instances are intended to reduce the time required to train large AI models, particularly those used in natural language processing and recommendation systems. These instances offer significant improvements in performance for AI inference and AI model training compared with previous-generation GPUs.

Customers using NVIDIA H100 Tensor Core GPUs can experience up to a 30x improvement in AI inference performance and up to 4x better performance in AI model training compared with the previous generation of NVIDIA A100 Tensor Core GPUs.

Customers can also leverage OCI Supercluster to connect multiple NVIDIA H100 GPUs for high-performance computing. These instances are expected to be available in the Oracle Cloud London Region and Oracle Cloud Chicago Region, with additional regions to follow.

NVIDIA L40S GPU Instances on OCI

These instances are designed for AI inferencing and training smaller to medium-sized AI models. These are expected to become available within the next year.

These instances offer improved performance for generative AI workloads and model fine-tuning compared with previous-generation GPUs. They have been tested to deliver up to a 20% improvement in performance for generative AI workloads and up to a 70% improvement in fine-tuning models compared with the previous generation of NVIDIA A100 GPUs.

Oracle to Use Ampere's Newest Chips in Its Cloud Offering

ORCL has announced its strategic decision to integrate Ampere’s flagship processor chips into its cloud computing service, providing a significant boost to Ampere.

Ampere Computing, founded by former Intel (INTC - Free Report) executives, specializes in the design of chips tailored for data centers, utilizing technology from Arm Holdings. The production of these advanced chips is outsourced to Taiwan Semiconductor Manufacturing Company, a global semiconductor manufacturing firm.

While Ampere has been a pioneer in creating energy-efficient processors for data centers, Intel and Advanced Micro Devices (AMD - Free Report) have responded by introducing its own energy-efficient solutions to compete in this growing market segment.

Intel introduced its Lakefield processors, which use a hybrid architecture combining high-performance cores with power-efficient cores, and its Alder Lake processors, which continue this approach with a big.LITTLE architecture. AMD introduced its EPYC server processors based on the Zen architecture, which offers competitive performance per watt in data center applications.

Notably, Oracle holds a substantial investment in Ampere and was an early adopter of AMD’s chips in 2021. Oracle Database, the world's leading database management system, has been fully enabled on Ampere CPUs. This Zacks Rank #3 (Hold) company is providing customers around the world the option to choose Ampere for performance and efficiency. You can see the complete list of today’s Zacks #1 Rank (Strong Buy) stocks here.

OCI Compute A2 Instances Powered by Ampere’s AmpereOne CPUs

These instances are anticipated to offer leading price-performance and boast a high core count, with up to 320 cores in the bare metal shape and up to 156 cores in the flexible VM shape. They are suitable for a variety of general-purpose cloud workloads, including web servers, video transcoding and CPU-based AI inference.

The high core count allows for enhanced performance, virtual machine density and scalability. Customers can also choose flexible VM shapes to optimize resource utilization and costs. These instances are planned to be introduced in the future.

By collaborating with NVIDIA and Ampere Computing, ORCL seeks to democratize access to cutting-edge GPU and CPU technology on its cloud infrastructure. This announcement positions Oracle as an early leader in providing cloud compute offerings tailored to support AI development and use.

Published in