CEVA-MM3000 utilizes a heterogeneous, multi-core architecture. The core of the platform consists of two specialized processors, a Stream Processor and a Vector Processor, combined into a complete multi-core system, including local and shared memories, peripherals, DMA and standard bridges to external busses. This comprehensive multi-core platform was designed specifically to meet low-power requirements for mobile devices.

The Stream Processor is responsible for bit stream coding, bit stream manipulation, and control code execution, while the Vector Processor performs filtering and vector type operations typical for pixel processing. Both the Stream and the Vector processors are designed for ease of programmability, and are supported by an optimizing C-compiler and application profiler. A set of pre-optimized video kernels for both processors is available to licensees, enabling the flexibility to combine efficient low-level code with rapid C-level development.

The platform delivers full-HD, 1080p video decode and encode at up to 60 frames per second (fps), supporting advanced video standards, including H.264, VC1, Real Video, AVS, VP6/7/8, as well as future standards such as SVC (Scalable Video Codec) and MVC (Multi-view Video Codec, for 3D video). CEVA-MM3000 also readily supports both video and image pre/post processing algorithms such as scaling, stabilizer, and noise reduction in addition to many proprietary add-on applications, for example, augmented reality, gesture recognition and face detection. Crucially, these additional algorithms and proprietary applications enabled by CEVA-MM3000 allow designers to reach new levels in multimedia innovation and performance, while offloading and even eliminating engines normally required to handle these complex features in a multimedia SoC.

Scalable and Programmable to enable Vast Array of Optimized Applications

To efficiently meet the precise requirements of a wide range of portable multimedia and home entertainment applications, CEVA-MM3000 utilizes a scalable and heterogeneous architecture. The number of Stream and Vector Processors in addition to the type and size of the memories can be configured to address each specific use-case. This scalability enables CEVA-MM3000 to deliver the optimal balance of power and performance for full-HD requirements in Smartphones, PMPs, Tablets and other mobile devices, through to Blu-ray DVD players, HDTVs, game consoles and multi-stream devices.

As a fully programmable engine, CEVA-M3000 readily supports any codec or other video-processing software. This programmability enables SoC designers to quickly adopt new video standards through software upgrades, requiring no changes to the hardware, and also easily develop proprietary video and imaging applications, including pre- and post-processing and video analytics. The resulting flexibility extends product life-cycle, enables design re-use and allows customers to differentiate their product offering with non-standard imaging or proprietary algorithms.

CEVA-MM3000 Platform

CEVA-MM3000 Features

	Industry's Highest Performance HD Video and Image Processing Platform for Connected Devices Built to support most advanced video codecs and profiles, including H.264, VC1, Real Video, AVS, VP6/7/8, MVC (3DV) up to 1080p 60fps Architecture ready for future standards including SVC and H.265 Encoding of H.264 HP at 1080p using the same platform
	Heterogeneous, Multi-core Architecture Optimized for Video and Image Signal Processing Stream Processor - Responsible for bit stream coding, bit stream manipulation, and control code execution Vector Processor - Deep SIMD processor for parallel filtering and pixel processing
	Scalable & Configurable Same architecture scaling from 720p to 1080p 60fps, from 6Mbps to 50Mbps Number of vector processors and number and type of stream processors is scalable Configurable memory size depending on the required performance Supports decoding of multiple streams, transcoding and video conferencing
	Multipurpose Platform, Supporting Various Video/Image Processing Needs Supports pre/post processing algorithms such as scaling, mirror, white balance, noise removal and DRC Supports Intelligent Video Processing including augmented reality, gesture recognition, face and object recognition
	Integrated software development tools Including compilers, multi-core debugger, simulators for each core and for the entire system Rich library of pre-optimized video kernels and filters
	Easily upgradeable to support any codec / profile SW from CEVA / 3rd parties / customers developments