Ampere Computing has unveiled its annual update, showcasing upcoming products and milestones that underscore its ongoing innovation in sustainable, ARM-based power-efficient computing for the Cloud and AI. This announcement was highlighted by the introduction of a groundbreaking CPU that promises to deliver 40% more performance than any current market offering.
Ampere Computing has unveiled its annual update, showcasing upcoming products and milestones that underscore its ongoing innovation in sustainable, ARM-based power-efficient computing for the Cloud and AI. This announcement was highlighted by the introduction of a groundbreaking CPU that promises to deliver 40% more performance than any current market offering.
The annual update included Ampere’s collaboration with Qualcomm Technologies. The two companies are working together to develop a joint solution for AI inferencing. This solution will leverage Qualcomm’s Cloud AI 100 inference solutions in conjunction with Ampere’s CPUs, addressing the increasing power demands of AI.
Ampere’s ARM approach focuses on maximizing performance while maintaining low power consumption, challenging the traditional notion that high performance requires high power. The company has demonstrated that superior performance can be achieved within an efficient power envelope, surpassing the capabilities of legacy CPUs.
Ampere emphasized the unsustainable trajectory of current power consumption trends in data centers. They advocate for retrofitting air-cooled environments with more efficient computing solutions and building new, environmentally sustainable data centers that align with the available power grid.
Other announcements during this year’s update included the following news highlights:
- Ampere Computing announced several significant developments in their latest update. The company is collaborating with Qualcomm Technologies to create a joint solution featuring Ampere CPUs and Qualcomm Cloud AI 100 Ultra, aimed at handling LLM inferencing on the industry’s most significant generative AI models.
- The update highlighted the expansion of Ampere’s 12-channel platform with the upcoming 256-core AmpereOne CPU. This CPU promises over 40% more performance than any current market CPU while maintaining the same air-cooled thermal solutions as the existing 192-core AmpereOne. The 192-core version is still expected to be released later this year.
- In another advancement, Meta’s Llama 3 is now operational on Ampere CPUs on Oracle Cloud. Performance data indicates that running Llama 3 on the 128-core Ampere Altra CPU, without a GPU, matches the performance of an NVIDIA A10 GPU paired with an x86 CPU, using only a third of the power.
- Ampere also announced the formation of a UCIe working group as part of the AI Platform Alliance. This initiative will enhance the flexibility of their CPUs by utilizing open interface technology, allowing the integration of other customers’ IP into future CPUs.
- Further details on AmpereOne performance and OEM/ODM platforms were provided, emphasizing its continued leadership in performance per watt, surpassing AMD Genoa and Bergamo by 50% and 15%, respectively. Data centers seeking to upgrade and consolidate infrastructure will be able to leverage AmpereOne’s up to 34% increase in performance per rack. New AmpereOne OEM and ODM platforms are expected to ship within a few months.
- In collaboration with NETINT, Ampere revealed a joint solution using Quadra T1U video processing chips and Ampere CPUs to transcode 360 live channels and provide real-time subtitling for 40 streams in multiple languages using OpenAI’s Whisper model.
- Lastly, Ampere has introduced a new FlexSKU feature, which allows customers to use the same SKU for both scale-out and scale-up use cases. This feature complements existing features like Memory Tagging, QOS Enforcement, and Mesh Congestion Management.
Ampere ARM-based CPUs are versatile and capable of running a wide range of workloads, from popular cloud-native applications to AI-driven tasks. This includes integrating AI with traditional applications such as data processing, web serving, and media delivery.
Engage with StorageReview
Newsletter | YouTube | Podcast iTunes/Spotify | Instagram | Twitter | TikTok | RSS Feed
