Full-Time
Sr. Principal Architect: AI Accelerator
Posted on 8/20/2024
Ampere
1,001-5,000 employees
High-performance processors for cloud computing
Compensation Overview
$163.1k - $271.9kAnnually
Senior
Company Historically Provides H1B Sponsorship
Santa Clara, CA, USA + 2 more
More locations: Durham, NC, USA | Portland, OR, USA
Hybrid role requiring in-office presence.
- BS degree in Electrical Engineering, Computer Engineering, or Computer Science and 12 years of experience; or MS degree and 8 years of experience; or PhD degree and 5 years of experience
- High-level knowledge and understanding of AI algorithms, compute, and memory requirements
- Experience architecting and/or designing high-performance domain specific accelerators, especially for AI or GPU applications
- Knowledge and understanding of: Cache-coherency protocols, Cache architecture and/or design, Virtual memory architecture, AMBA suite of interface protocols, PCIe and CXL protocols
- Experience with implementing ARMv9, Nvidia and/or AMD GPU ISAs, especially vector and matrix instructions and pipeline-based micro-architectures
- Experience with PCIe host/device interface models, IO virtualization and DMA engines
- Experience with performance modeling, e.g. in Python or C++
- Must be self-driven, curious, organized and comfortable with ambiguity
- Ability to learn and adapt, one of the foundational principles of Ampere
- Great communication skills, comfortable with reaching out and asking questions, and brainstorming with peers
- Contribute to the creation of a groundbreaking AI compute product that supports sustainability and performance, designed to operate efficiently across a variety of data center infrastructures
- Be a critical technical leader on our new AI product roadmap, leveraging experience and expertise to shape the architecture, evaluate trade-offs and make critical decisions
- Work closely with the SW team and other HW architects and designers to define how target workloads will be mapped to domain-specific HW functions, including the host-device interface, data movement, and vector and matrix execution pipelines
- Help define a flexible HW architecture that meets the performance targets across different form-factors, configurations, and TDPs
- Work with the driver team to help define the host/device interface, including command structures and DMA engines
- Provide guidance and leadership to other architects on the product, to ensure the overall product works efficiently end-to-end for the target workloads
- Help with performance modeling and analysis of the HW for targeted workloads
- Create and maintain clear, detailed and comprehensive product architecture specifications
Ampere Computing specializes in high-performance processors for cloud applications, offering models that outperform competitors by significant margins. Their processors are designed for sustainability, utilizing a unique architecture that maximizes performance while minimizing power consumption. The product lines, including the Ampere Altra Family and AmpereOne Family, are scalable for various data center environments and edge computing needs. Ampere also supports developers with tools and resources to create cloud-native solutions.
Company Size
1,001-5,000
Company Stage
Private
Total Funding
$369.6M
Headquarters
Santa Clara, California
Founded
2017
Simplify's Take
What believers are saying
- Ampere's expansion into telecom aligns with the growing demand for 5G infrastructure.
- Partnership with SpaceTech highlights interest in sustainable computing for smart cities.
- Potential acquisition by Arm Holdings could enhance Ampere's market reach and resources.
What critics are saying
- AWS's 896-core instance challenges Ampere's position in high-performance computing.
- Potential acquisition by Arm suggests strategic vulnerabilities for Ampere.
- Rising AI accelerator prices may impact Ampere's cost competitiveness.
What makes Ampere unique
- Ampere's processors offer 50% higher performance than Genoa and 68% than Bergamo.
- Their single-threaded core architecture ensures high performance with lower power consumption.
- Ampere's scalable design supports diverse environments from data centers to edge deployments.
Help us improve and share your feedback! Did you find this helpful?
Growth & Insights and Company News
6 month growth
↓ -1%1 year growth
↑ 0%2 year growth
↑ 0%Ampere is accelerating its expansion from Arm-based server processors for AI processing into networking chips for the telecom market. The company unveiled the ampere Altra Family of processors uniquely suited to address growing marketAmpere announced today that it is accelerating its effort to address the telecom market, bringingits high-performance, energy-efficient processors to next-generation RAN networks. With the rapid expansion of 5G, edge computing and AI-driven workloads, the company said it is seizing a major market opportunity by enabling telecom providers to meet growing performance demands while lowering costs and energy use.According to Grandview Research, the global telecom services market size is estimated to reach $1.9 billion in 2024 and is projected to grow at a compound annual growth rate (CAGR) of 6.5% from 2025 to 2030. “Rising spending on the deployment of 5G infrastructures due to the shift in customer inclination toward next-generation technologies and smartphone devices is one of the key factors driving this industry,” Grandview said.Ampere said its processors provide numerous advantages for the telecom market due to their high core density, energy efficiency and scalability, which support the industry’s growing need for high-performance and low power. Low power consumption and efficient thermal design make Ampere-based platforms ideal for a variety of edge form factors found in telecom network infrastructure.They help lower operational costs, a key benefit for telcos managing large-scale deployments across many regions with unpredictable and rising energy costs.Because of this strong product fit, Ampere is strengthening its presence in the telecom sector to capitalize on this significant market opportunity. The company views this market as a natural extension of the Cloud Native workloads it already supports today, delivering Ampere’s benefits to the edge and across distributed networks.Securing new customersToday, Ampere is announcing new trials with global telecom customers as it expands its reach inthis market.Ampere and Parallel Wireless are collaborating on the deployment of the Parallel WirelessGreenRANTM Hardware-Agnostic cloud-native Distributed Unit (DU) stack for 5G SA, 5G NSA,4G and 2G, which will be the first stack to commercialize a comprehensive portfolio ofcellular network technologies on Ampere
To improve performance, energy efficiency, and sustainability, SpaceTech partnered with Ampere Computing to explore a transition from x86 edge servers to Ampere Altra family Arm-based servers.
SoftBank and Arm Holdings are said to be in talks to acquire Ampere Computing.
SoftBank's Arm Holdings is exploring a potential acquisition of Ampere Computing, an Oracle-backed semiconductor designer valued at $8 billion in a 2021 SoftBank investment proposal. Ampere, which designs chips using Arm's technology, is considering strategic options amid increased competition in the AI chip market. Talks are ongoing, but a deal is not guaranteed. Ampere's interest in a sale suggests challenges in pursuing an IPO. Arm's acquisition could enhance its push into the data center market.
Arm Holdings PLC - ADR ARM+0.08%Get Free Report is exploring the potential acquisition of semiconductor designer Ampere Computing LLC, signaling further consolidation in the artificial intelligence chip sector.