‘For many AI applications, GPUs are compute overkill, consuming much more power and money than needed’: How Ampere Computing plans to ride the AI wave

Key Takeaways:

– Ampere Computing is a startup challenging the dominance of tech giants in the computing industry.
– They aim to address the increasing demand for computing power with a low-power, high-performance solution.
– Their offering has been adopted by major hyperscalers worldwide.
– Ampere has broken through the scaling wall multiple times with its CPUs and plans to continue scaling in ways that legacy architectures cannot.
– The company believes that the industry needs something new and their compute solution is filling that gap.
– They see increased adoption in the enterprise as companies look to optimize their data center footprint.
– Ampere’s strategy is to develop sustainable processors using the best available technologies.
– Their focus is on releasing CPUs that are more efficient with higher core counts, memory bandwidth, and IO capabilities.
– They serve a broad spectrum of applications and workloads, including AI inference, web services, databases, and video processing.
– Ampere CPUs deliver higher performance at lower power consumption compared to legacy x86 processors.
– They work with major hyperscalers across the US, Europe, and China, as well as OEMs in the enterprise market.
– Ampere’s CPUs are also suitable for edge deployments with stringent space and power requirements.
– AI will continue to be a major topic of conversation, but the focus will shift from training to deploying AI inference.
– Achieving scale in AI inference will be limited by performance, cost, and availability, making low-power, high-performance CPUs like Ampere’s attractive.
– Sustainability and energy efficiency will become even more important in the context of AI.
– Ampere’s products provide a more efficient and cost-effective alternative to GPUs for many AI applications.
– They save power, space, and cost when running AI workloads alongside other workloads like databases or web services.

TechRadar:

Ampere Computing is a startup that’s making waves in the tech industry by challenging the dominance of tech giants like AMD, Nvidia, and Intel. With the rise of AI, the demand for computing power has skyrocketed, along with energy costs and demand on power grids. Ampere aims to address this with a low-power, high-performance solution.

Despite being the underdog, Ampere’s offering has been adopted by nearly all major hyperscalers worldwide. It has broken through the scaling wall multiple times with its CPUs, and the company plans to continue scaling in ways that legacy architectures can’t. We spoke to Ampere CPO Jeff Wittich about his company’s success and future plans.

I feel sometimes that challenger startups, like Ampere Computing, are stuck between a rock and a hard place. On one side, you’ve got multi billion dollar companies like AMD, Nvidia and Intel and on the other hand, hyperscalers like Microsoft, Google and Amazon that have their own offerings. How does it feel to be the little mammal in the land of dinosaurs?

Source link

AI Eclipse TLDR:

Ampere Computing is a startup that is disrupting the tech industry by challenging major players like AMD, Nvidia, and Intel. With the increasing demand for computing power due to the rise of AI, Ampere aims to address this by offering a low-power, high-performance solution. Despite being the underdog, Ampere’s products have been adopted by major hyperscalers worldwide and the company plans to continue scaling in ways that legacy architectures cannot.

In an interview with Ampere’s CPO, Jeff Wittich, he expressed excitement about the company’s success and the need for a new compute solution in the cloud. The industry needs something new rather than more of the same from established players. Ampere has rapidly grown and gained adoption from big hyperscalers and enterprises who see the value in its low-power, high-performance solution.

When it comes to physical cores, Ampere has been the leader in high core count in the server CPU market. While competitors like AMD and Intel are catching up, Ampere has been able to break through the scaling wall multiple times by taking a new approach to CPU design. This allows them to continue scaling in ways that legacy architectures cannot.

Another potential threat to Ampere is the rise of RISC-V, a microarchitecture supported by China. However, Ampere’s core strategy is to develop sustainable processors using the best available technologies. While they remain open to using RISC-V if it aligns with their goals, they will choose technologies that can be easily used by their customers.

As for Ampere’s future plans, they will focus on releasing CPUs that are more efficient, deliver higher core counts, and provide more memory bandwidth and IO capabilities. This will allow them to meet the increasing demands of important workloads like AI inferencing while also meeting sustainability goals. They will also introduce additional features that provide flexibility for cloud providers to meet customer applications.

Ampere’s CPUs serve a broad spectrum of applications and have been adopted by major hyperscalers across the world. They offer 2x the performance for workloads like AI inference, web services, and databases at half the power of legacy x86 processors. Ampere CPUs are also suitable for edge deployments with stringent space and power requirements.

In terms of the AI market, Ampere believes AI will continue to be an important topic of conversation. However, the conversation will shift from training neural networks to deploying them, also known as AI inference. CPUs, particularly low-power, high-performance CPUs like Ampere’s, will become more attractive for efficient and cost-effective execution of AI inference models. Sustainability and energy efficiency will also become more important as data centers struggle to meet the increasing demand for computing power.

Ampere is addressing the AI market opportunity by offering CPUs that are more efficient and cost-effective for most AI applications. They believe that GPUs are often overkill for most inferencing workloads and replacing them with CPUs can save power, space, and cost. This trend is already becoming evident in the industry.