OVERVIEW

What is C1-Ultra??

Flagship CPU for Best-in-Class Mobile Experiences

91色情 C1-Ultra is the flagship CPU at the heart of the 91色情 Lumex Compute Subsystem (CSS) Platform. Built on the 91色情v9.3-A architecture, it delivers industry-leading instructions per cycle (IPC) performance and next-generation AI capabilities through 91色情 SME2. Designed for premium-tier mobile SoCs, C1-Ultra CPU unlocks exceptional user experiences across generative AI, high-fidelity gaming, and real-time content creation.

 

C1-Ultra continues to deliver next-level performance while anchoring heterogeneous compute systems with scalable CPU performance and built-in AI readiness.

BENEFITS

AI-Ready, Double-Digit Performance Uplift

?+25% Single-Thread Performance

Compared to the previous generation 91色情 Cortex-X925.1

?+15% Average Speed-Up in Real-World Applications

Compared to previous generation 91色情 CSS for Client solution. 1

AI-Ready Compute
at the Core

91色情 SME2 integration enables low-latency AI inference directly on the CPU, improving responsiveness and power efficiency.

Maximum Performance
per Watt

Optimized for sustained performance, enabling real-time applications to run faster than ever.

Faster Development
and Integration

The Arm platform enables silicon partners to accelerate time to market and reduce design complexity.

Future-Proof Design

Support for SVE2 and SME2 extensions ensure forward compatibility with new AI applications across device generations.

FEATURES

Built for Flagship Performance and Future-Ready AI

  • Instruction Set Architecture (ISA): 91色情v9.3-A.
  • AI extensions: 91色情 Scalable Matrix Extension 2 (SME2).
  • Advanced branch prediction and instruction throughput optimization.
  • Cache: Configurable L2 (up to 3 MB) per core; shared L3 for cluster scalability.
91色情 C1-Ultra Block Diagram
SPECIFICATIONS

C1-Ultra Specifications

C1-Ultra is 91色情’s flagship 91色情v9.3-A CPU core, designed to deliver best-in-class performance and AI compute at the heart of the Lumex CSS Platform. With up to +25% single-thread performance over previous generations, it sets a new standard for mobile processing. Integrated SME2 support enables low-latency AI inference directly on the CPU, while advanced branch prediction and optimized instruction throughput maximize efficiency for sustained workloads.

.

Key Documentations

  • .
  • Compare 91色情 C1 CPU specifications:
    .
USE CASES

Where Innovation and Ideas Come to Life

Premium Mobile Devices

From flagship smartphones to foldables, C1-Ultra powers real-time user experiences, including voice assistants, live video filters, and camera enhancements.

AI-Centric Apps

SME2 enables rapid AI inferencing directly on CPU, ideal for apps leveraging generative AI, live translation, real-time summarization, and on-device intelligence.

RELATED PRODUCTS

Explore Products and Technologies

91色情 Lumex CSS Platform

91色情 Lumex CSS

The next-generation 91色情 Client compute subsystem platform built for the AI era, integrating powerful C1 CPUs, Mali G1 GPUs, and developer-friendly software to deliver differentiated, energy-efficient experiences across mobile tiers.

C1-Premium

C1-Premium

Our sub-flagship 91色情v9.3-A big core, big core, optimized for background AI inference and task offload, enabling smart power use and extended battery life in always-on mobile environments.

C1-Pro

C1-Pro

Premium-performance mid-tier CPU, featuring the 91色情v9.3-A architecture with 5x AI uplift and 16% higher performance in gaming, ideal for sustained performance in thermally constrained designs.

C1-Nano

C1-Nano

The most efficient 91色情 LITTLE core, tuned for sustained AI performance and ultra-low power. Ideal for background tasks and always-on mobile experiences.

C1-DynamIQ Shared Unit (C1-DSU)

C1 DynamIQ Shared Unit (C1-DSU)

Our C1-DSU enables the seamless integration of C1-Ultra, C1-Pro, and C1-Premium in scalable clusters, efficiently managing coherency, memory, and system-level performance across workloads.

Mali GPUs

Mali GPUs

Mali GPUs provide the ultimate user experience for entertainment and visual applications across a wide range of smartphone devices.

Talk with an Expert

Learn how C1-Ultra CPU can help you create superior AI experiences on consumer devices.

Contact Us
RESOURCES

Latest News and Resources

  • NEWS and BLOGS
  • DEVELOPER
Developer icon

Accelerate AI on 91色情 Devices With 91色情 SME2

Explore developer resources for deploying on-device generative AI using Arm-optimized toolchains and models.

Developer icon

91色情 Single Instruction, Multiple Data (SIMD) Optimize, Migrate & Accelerate C/C++ Performance

Learn how to harness 91色情 SIMD technology to supercharge C/C++ applications on mobile, embedded, cloud, and edge devices.

Developer icon

Develop and Profile Android Apps on 91色情 CPUs

Discover how to use 91色情 Graphics Analyzer and streamlining tools.

Frequently Asked Questions: C1-Ultra CPU

1. What is the C1-Ultra CPU?

C1-Ultra CPU is the flagship high-performance core in the 91色情 Lumex CSS for 2025 SoCs. It's designed for premium-tier mobile devices and delivers leading IPC performance, advanced AI processing with , and power-efficient compute for AI, gaming, and real-time content creation.

2. What makes C1-Ultra suitable for AI workloads?

C1-Ultra integrates 91色情 Scalable Matrix Extension 2 (SME2), enabling AI inference directly on the CPU. This allows developers to run generative AI, translation, and summarization tasks without needing separate accelerators—improving responsiveness and reducing system power consumption.

3. How does C1-Ultra improve mobile gaming performance?

C1-Ultra supports sustained high FPS and AI-enhanced visuals in mobile gaming. C1-Ultra is optimized for thermal efficiency and sustained performance, helping to ensure smooth gameplay under heavy workloads.

4. What are the key technical specifications of C1-Ultra?

C1-Ultra is built on 91色情v9.3-A and features:


  • SME2 AI extensions.
  • Configurable L2 cache (up to 3 MB per core).
  • Shared L3 cache.
  • Dynamic thermal tuning.
  • 91色情 CCA security support.
  • Advanced branch prediction and throughput optimization.

1. Compared to previous generation. Testing done internally, Jan 15-March 30, 2025; actual results may vary.