The block diagram of the EyeQ4 reveals a deliberate separation of general-purpose compute and vision-specific pipelines.
The EyeQ4 is not programmable via standard C compilers for the vision accelerators. Instead, Mobileye provides the EyeQ Software Development Kit (SDK) which includes:
Note: The datasheet explicitly states that third-party CUDA or OpenCL code is not supported.
While the full datasheet includes hundreds of timings, key electrical parameters for integration are:
The EyeQ4 is purpose-built for Level 2+ and Level 3 autonomy. Unlike general-purpose SoCs, it is a heterogeneous multi-processor unit with hardware accelerators dedicated to computer vision.
| Part Number | Description | |------------------|------------------------------------| | EYQ4-8C2-1 | EyeQ4, 8GB LPDDR4, industrial temp | | EYQ4-4C1-A | EyeQ4, 4GB LPDDR4, automotive temp | (Note: Check official pricing/ordering with Intel/Mobileye representatives)
Unlike a general-purpose GPU, the EyeQ4 features fixed-function accelerators that offer massive power savings:
The Mobileye EyeQ4 Go to product viewer dialog for this item.
is a high-performance System-on-Chip (SoC) designed for Advanced Driver Assistance Systems (ADAS) and autonomous driving. Manufactured by STMicroelectronics using 28nm FD-SOI technology, it provides 10x the processing power of its predecessor, the EyeQ3, while maintaining a low power envelope. Technical Specifications
The EyeQ4 architecture utilizes a heterogeneous mix of specialized accelerators to achieve high efficiency. Specification Performance 2.5 TOPS (High variant) / ~1.1 TOPS (Mid variant) Power Consumption ~3 Watts (Automotive grade) CPU Cores 4 multi-threaded MIPS InterAptiv cores (4 threads each) Vision Accelerators
6 Vector Microcode Processors (VMP), 2 Multithreaded Processing Clusters (MPC), 2 Programmable Macro Arrays (PMA) Camera Support Up to 8 cameras simultaneously at 36 fps Safety Standard ISO 26262 compliant; ASIL-B(D) level Package Flip-Chip FBGA 784-pin (22.5 x 22.5 x 1.7mm) Key Capabilities The Evolution of EyeQ - Mobileye
Mobileye EyeQ4 is a high-performance vision processor System-on-Chip (SoC) designed specifically for Advanced Driver Assistance Systems (ADAS) and semi-autonomous driving. Launched in 2018, it represented a significant jump in performance, offering roughly 10 times the processing power of its predecessor, the EyeQ3, with only a 20% increase in power consumption. Core Specifications & Performance
The EyeQ4 architecture is built on a 28nm FD-SOI (Fully Depleted Silicon On Insulator) process technology from STMicroelectronics, which is critical for its high efficiency. Specification Details Performance Over 2.5 Teraflops (TFLOPS) or 2.5 TOPS Power Consumption Approximately 3 Watts for typical automotive use Video Processing
Can process up to 8 cameras simultaneously at 36 FPS (high version) Package Size 22.5 x 22.5 mm (784-pin Flip-Chip FBGA) Architecture Overview
The EyeQ4 uses a heterogeneous "accelerator-rich" architecture. It doesn't rely solely on standard CPUs but instead uses four specialized classes of programmable accelerators designed for computer vision tasks.
MIPS CPU Cores: Features four multi-threaded MIPS InterAptiv processor cores (with 4 threads each) for general-purpose management and control.
Vector Microcode Processors (VMP): Six programmable cores optimized for vision algorithms like image filtering and feature extraction.
Multithreaded Processing Cluster (MPC): Two cores designed to be more versatile than a GPU and more efficient than a CPU for complex vision tasks.
Programmable Macro Array (PMA): Two cores offering compute density similar to fixed-function hardware while remaining fully programmable. Key ADAS Capabilities
The EyeQ4 enables several safety and autonomous functions that were difficult for previous generations: The Evolution of EyeQ - Mobileye
Deep Learning Accelerator. Dedicated high-performance AI engine. The main source of horse power for convolutional neural networks. ZF and Mobileye Safety Technology Chosen by Toyota
Mobileye EyeQ4 is an automotive-grade vision processor (SoC) designed by Mobileye and manufactured by STMicroelectronics using 28nm FD-SOI technology. It represents a massive leap in processing power for Advanced Driver Assistance Systems (ADAS) compared to its predecessors. Core Specifications Architecture
: High-performance multi-core design including 4 multi-threaded MIPS InterAptiv CPU cores, 6 Vector Microcode Processors (VMP), 2 Multithreaded Processing Clusters (MPC), and 2 Programmable Macro Arrays (PMA). Performance : Delivers 2.5 Teraflops (TOPS)
of processing power, which is roughly 10x the capability of the EyeQ3. Efficiency : Consumes approximately
, making it highly energy-efficient relative to its output (only 20% more power than the EyeQ3 for 10x the power). Vision Input : Supports visual input from up to simultaneously at 30fps. Key Capabilities Advanced Detection
: Includes vehicle detection from any angle, next-generation lane detection, and traffic light detection. Environmental Modeling
: Capable of full environmental modeling and holistic path planning. : Supports Mobileye's Road Experience Management (REM) for crowd-sourced high-definition mapping. Safety Features
: Powers Automated Emergency Braking (AEB), Forward Collision Warning (FCW), and Adaptive Cruise Control (ACC). Review: The "Sweet Spot" for Semi-Autonomous Driving
The EyeQ4 is widely considered the processor that moved ADAS from simple "passive" alerts to "active" semi-autonomous driving. Unmatched Efficiency
: At just 3W, it delivers heavy-duty processing without requiring complex cooling systems, a critical factor for automotive reliability. Massive Scalability eyeq4 datasheet
: It was designed to support everything from basic mono-camera systems to complex "Tri-cam" setups found in luxury brands like BMW. Proven Reliability
: Already integrated into over 160 car models from major OEMs like GM, Nissan, and Honda. Generationally Older : While powerful, it has since been surpassed by the EyeQ Ultra
, which offer significantly higher TOPS for Level 4/5 autonomy. Closed System
: Historically, Mobileye chips have been more "black box" systems, though later generations (like EyeQ5) began moving toward more open software platforms. performance against the newer
The Mobileye EyeQ4 is a high-performance vision system-on-chip (SoC) designed for Advanced Driver Assistance Systems (ADAS) and semi-autonomous driving. It provides approximately 2.5 teraflops of processing power while maintaining a low-power automotive-grade envelope of roughly 3W. Technical Specifications Summary
The EyeQ4 architecture is based on a heterogeneous computing model that assigns specific tasks to specialized cores for maximum efficiency. Feature Specification Details Processor Cores
4x multi-threaded 64-bit RISC MIPS CPUs (4 hardware threads each) Vision Accelerators
6x Vector Microcode Processors (VMP), 2x Multithreaded Processing Clusters (MPC), 2x Programmable Macro Arrays (PMA) Compute Power >2.5 Teraflops (or 2.5 TOPS depending on variant) Power Consumption ~3 Watts (up to 5W in some high-load configurations) Process Node
28nm Fully Depleted Silicon On Insulator (FD-SOI) by STMicroelectronics Camera Support Up to 8 cameras simultaneously at 36 FPS Safety Standard ISO 26262 compliant with ASIL-B(D) safety level Packaging Flip-Chip FBGA 784-pin; 22.5 x 22.5 x 1.7 mm EyeQ4 Variant Differences
Mobileye developed multiple versions of the chip to support different vehicle capabilities: EyeQ4-High (EyeQ4H)
: The most capable version, supporting trifocal front-sensing, surround-view systems (4 cameras), and sensor fusion with radar and laser scanners. EyeQ4-Medium (EyeQ4M)
: A cost-optimized variant with a subset of cores, typically used for monocular or trifocal camera configurations in standard ADAS applications. Key Interfaces and Connectivity
According to the EyeQ4 Product Brief, the chip includes the following I/O: Memory: Dual 32-bit LPDDR4 SDRAM interfaces at 1.6GHz. Network: 1Gb Ethernet port.
Video Input: 4x MIPI CSI-2 Rx serial video ports and 1x parallel video port.
Automotive Buses: 3x CAN ports (>1Mbps), 3x UART, 3x I2C, and 4x SPI interfaces. Documentation and Resources Mobileye EyeQ4 Vision Processor Family - Yole Group
EyeQ4 Datasheet Write-up
The EyeQ4 is a high-performance, low-power System-on-Chip (SoC) designed for advanced driver-assistance systems (ADAS) and autonomous driving applications. Developed by Mobileye, a leading provider of computer vision and machine learning technologies, the EyeQ4 is a fourth-generation SoC that offers significant improvements in processing power, memory, and software capabilities compared to its predecessors.
Overview
The EyeQ4 datasheet provides an in-depth look at the SoC's architecture, features, and specifications. Here are some key highlights:
Key Features
The EyeQ4 datasheet highlights several key features that make it an attractive solution for ADAS and autonomous driving applications:
Applications
The EyeQ4 is designed for a range of ADAS and autonomous driving applications, including:
Conclusion
The EyeQ4 datasheet provides a comprehensive overview of Mobileye's latest SoC for ADAS and autonomous driving applications. With its powerful processing capabilities, large memory capacity, and support for a range of sensors and software frameworks, the EyeQ4 is well-suited for demanding applications like computer vision, machine learning, and autonomous driving. As the automotive industry continues to evolve towards more advanced driver-assistance systems and autonomous vehicles, the EyeQ4 is poised to play a key role in enabling these technologies.
The Mobileye EyeQ4 is a high-performance vision-based System-on-Chip (SoC) designed specifically for Advanced Driver Assistance Systems (ADAS) and autonomous driving. Leveraging a heterogeneous multi-core architecture, it delivers significant leaps in computational efficiency compared to its predecessors.
The following data summarizes the key specifications and architectural details typically found in an EyeQ4 datasheet. Core Performance Specifications
The EyeQ4 is engineered to provide "super-computer" capabilities within a strictly constrained automotive power envelope.
Computational Power: Up to 2.5 Tera Operations Per Second (TOPS). The block diagram of the EyeQ4 reveals a
Power Consumption: Approximately 3 Watts, achieved through a high-efficiency 28nm FD-SOI (Fully Depleted Silicon On Insulator) manufacturing process.
Safety Rating: Designed to meet ISO-26262 standards with a safety level of ASIL-B(D).
Camera Support: Capable of processing up to 8 cameras simultaneously at 36 frames per second (fps). Architectural Overview
The EyeQ4 utilizes a diverse set of specialized accelerators to handle complex computer vision and deep learning tasks efficiently. Description CPU Cores General Purpose Compute
Quad-core MIPS interAptiv processors with multi-threading (up to 4 threads per core). VMP Vector Microcode Processor
6 cores dedicated to VLIW and SIMD operations, ideal for short integral types in vision algorithms. MPC Multi-threaded Processor Cluster
2 cores offering higher efficiency than standard CPUs and more versatility than a GPU. PMA Programmable Macro Array
2 Cores using a CGRA dataflow machine architecture for dense computer vision algorithms. EyeQ4 Family Variants
Mobileye offers different versions of the EyeQ4 to provide a scalable solution for varying levels of vehicle autonomy.
EyeQ4 High: The full-capability version designed for surround-view systems and trifocal front-sensing. It processes information from multiple cameras, radars, and lidars to create a "safety cocoon" around the vehicle.
EyeQ4 Mid: A subset version tailored for mid-range ADAS. It integrates fewer cores (e.g., three MIPS cores and four VMP cores) and is typically used in single-camera or trifocal configurations.
EyeQ4 Lite: Optimized for entry-level NCAP compliance and basic collision avoidance features. Key Features and Applications
The EyeQ4 datasheet highlights several next-generation ADAS capabilities:
Object Detection: Support for vehicle detection from any angle and pedestrian/cyclist identification.
Mapping (REM): Integration with Mobileye Road Experience Management (REM) for crowdsourced high-definition mapping.
Lane Detection: Next-generation lane and road boundary detection for centering and departure warnings.
Sensor Fusion: Efficiently fuses data from optical sensors with radar and scanning-beam lasers. Physical and Electrical Characteristics
Detailed hardware integration data for the EyeQ4-Mid and EyeQ4-High includes: Package: Flip-Chip FBGA with 784 pins. Dimensions: 22.5 mm x 22.5 mm x 1.7 mm.
Manufacturing: Produced by STMicroelectronics using a proprietary 28nm process.
Interfaces: Includes high-speed automotive interfaces such as Ethernet, CAN, and PCIe for ECU communication.
For developers seeking to integrate this chip, the Mobileye Technology Page provides further insights into the evolution of this architecture and its role in modern autonomous platforms.
The Mobileye EyeQ4 is a high-performance vision processor designed for Advanced Driver Assistance Systems (ADAS) and autonomous driving, offering a massive leap in processing power over its predecessors. Key Technical Specifications Performance: 2.5 Tera Operations Per Second (TOPS).
Efficiency: 10x more powerful than EyeQ3 with only a 20% increase in power consumption. Architecture:
Manufactured using 28nm FD-SOI technology by STMicroelectronics.
Features 14 computing cores, including specialized vector accelerators. Integrates four multi-threaded MIPS InterAptiv cores.
Camera Support: Capable of processing up to 10 cameras simultaneously at 36 frames per second. "Interesting" Breakthroughs & Capabilities
High Utilization: Achieves a 96% utilization rate, which is significantly higher than most general-purpose GPUs.
Complex Recognition: Supports "any-angle" vehicle detection and next-generation lane detection.
Scalability: Used in configurations ranging from a single "Mono" camera for collision avoidance to "Tricam" setups for semi-autonomous driving. Note: The datasheet explicitly states that third-party CUDA
Safety Standards: Designed for compliance with EU NCAP and US NHTSA regulatory requirements.
Market Impact: By 2018, it was already launched in 78 different vehicle models from 16 major manufacturers like BMW, Nissan, and GM.
💡 Pro-Tip: For specific implementation, designers often pair the EyeQ4 with a dedicated power management unit like the TI LP875761-Q1 to handle the SoC's core rail requirements.
If you'd like to dive deeper, would you prefer details on the programming architecture or its role in specific car models?
Mobileye EyeQ4 is a high-performance vision system-on-chip (SoC) designed specifically for advanced driver-assistance systems (ADAS) and semi-autonomous driving. Developed by Mobileye (an Intel company) and manufactured by STMicroelectronics, it represented a massive jump in processing power—roughly 10 times that of its predecessor, the EyeQ3, while consuming only 20% more power. Key Specifications & Architecture
is built on a 28nm Fully Depleted Silicon On Insulator (FD-SOI) process, which is critical for maintaining high performance with a low power envelope of approximately 3 watts. -High Specifications Compute Performance ~2.5 Teraflops (1.26 Billion MAC/s) CPU Cores
4 multi-threaded MIPS InterAptiv cores (4 hardware threads each) Vision Accelerators
6 Vector Microcode Processors (VMP), 2 Multithreaded Processing Clusters (MPC), 2 Programmable Macro Arrays (PMA) Camera Input Up to 8 cameras simultaneously at 36 fps Functional Safety ISO 26262 ASIL-B(D) Packaging Flip-Chip FBGA 784-pin (22.5 x 22.5 x 1.7mm) Specialized Processing Cores The "magic" of the
datasheet lies in its heterogeneous architecture, which uses different types of proprietary accelerators for specific vision tasks:
Vector Microcode Processors (VMP): A VLIW and SIMD machine optimized for computer vision and deep learning algorithms.
Multithreaded Processing Cluster (MPC): More versatile and efficient than a standard GPU, handling complex control and data management.
Programmable Macro Array (PMA): A CGRA dataflow machine providing compute density similar to fixed-function hardware while remaining programmable. Product Variants: High vs. Mid The Evolution of EyeQ - Mobileye
Inside the EyeQ4: The "Supercomputer" Driving Your Next Car Mobileye EyeQ4
isn't just another chip; it's the silicon brain that moved Advanced Driver Assistance Systems (ADAS) from simple warnings to near-autonomous "safety cocoons". Launched in 2018, this System-on-Chip (SoC) provides a staggering 10x more processing power
than its predecessor, the EyeQ3, while keeping power consumption remarkably low.
Here is a deep dive into the technical specifications and capabilities that make the a landmark in automotive technology. Technical Specifications: The Power of Efficiency is manufactured using 28nm Fully Depleted Silicon On Insulator (FD-SOI) technology by STMicroelectronics
. This specialized manufacturing process is what allows it to deliver "super-computer" performance within a tiny power envelope. Computational Performance: 2.5 Teraflops (trillions of operations per second). Power Consumption: Approximately , which is less than many standard mobile phone processors. Architecture: A heterogeneous mix of cores designed for specific tasks: Four multi-threaded MIPS cores. VMP (Vector Microcode Processors):
Six cores for image processing and integral types used in computer vision. MPC (Multi-threaded Processing Cluster):
Two cores more versatile than GPUs and more efficient than CPUs. PMA (Programmable Macro Array):
Two cores providing high compute density for dense computer vision algorithms. Supports dual 1.6GHz, 32-bit LPDDR4 SDRAM interfaces. Connectivity:
Includes a 1Gb Ethernet port, multiple CAN ports (>1Mbps), UART, and I2C interfaces. Safety Rating: Designed according to ISO-26262 to provide safety levels. Advanced ADAS Capabilities What does all that silicon power actually do? The is designed to process information from up to eight cameras simultaneously
at 36 frames per second. This allows it to support sophisticated features that were previously impossible for a single chip: ZF and Mobileye Safety Technology Chosen by Toyota
The Mobileye EyeQ4 is a 28nm FD-SOI, high-performance System-on-Chip (SoC) designed for camera-based Advanced Driver Assistance Systems (ADAS), delivering over 2.5 teraflops of processing power at 3W. Featuring six VMP cores, two MPC cores, and two PMA cores, it supports up to 8 simultaneous cameras for advanced computer vision and autonomous emergency braking. For more details, visit Mobileye.
EYEQ4 Datasheet Overview
The EYEQ4 is a highly integrated, low-power, and compact image signal processor (ISP) designed for various camera applications. The EYEQ4 datasheet provides detailed specifications, features, and technical information for the chip.
Key Features:
Datasheet Contents:
The EYEQ4 datasheet typically includes:
Target Applications:
The EYEQ4 is suitable for various camera applications, including:
| Parameter | Value | |-----------|-------| | Manufacturing Process | 28nm FD-SOI (Fully Depleted Silicon On Insulator) | | Peak Performance | 2.5 TOPS (Trillions of Operations Per Second) | | Power Consumption | ~3W (typical, automotive grade) | | Camera Inputs | Up to 8 cameras (raw image sensors via MIPI CSI-2) | | CPU Cores | 4× MIPS interAptiv (lockstep for safety) | | Vision Accelerators | 4× Vector Microcode Processors (VMPs) | | Memory Interface | LPDDR4 / DDR3L (up to 4 GB, 64-bit) | | Safety Level | ASIL B (system) / ASIL D (for some safety paths) | | Operating Temp | -40°C to +105°C (Tj) |