Today’s consumer devices feature rich, graphical user interfaces and run interactive applications like games and mobile web. Vivante GPU IP delivers high performance, high quality total graphics solutions at low power to drive stunning visual displays on mobile phones, MIDs and netbooks, digital cameras, portable navigation systems, personal media players, e-books, digital picture frames, automotive infotainment, set-top boxes, HDTV, and gaming systems. All Vivante GPUs support HD 1080p display resolution.
Hardware / Software / Ecosystem
An optimized software stack, complete software development tools, and a growing application ecosystem are supported by our robust graphics pipeline designed to industry standard API specifications for DirectFB, OpenGL ES 2.0, OpenGL ES 1.1, and OpenVG 1.1. And with full support for Android, Linux, and Windows embedded development platforms, every Vivante GPU core comes fully featured with console-quality graphics and video post-processing capabilities.
Silicon Efficiency
Our ScalarMorphicTM architecture and other innovations result in the smallest die size in every GPU class. Vivante solutions are optimized to reduce bandwidth and power, and to minimize the CPU load for real applications. These features are absolutely essential for building efficient embedded systems.
Full-featured high performance, low power, and smallest area mean that Vivante delivers the best in graphics quality at the lowest silicon cost.
Easy to Integrate
With support for free-running AXI/AHB or OCP interfaces, customers can easily integrate our cores into their systems. Vivante licensees quickly and consistently reach target frequency and optimal final silicon area using provided reference implementation kits for Cadence and Synopsys design tool environments.
Performance**
Core clock 250 MHz (65nm LP) to 375 MHz (65nm G+)
25M triangles per second (65nm LP) to 38M triangles per second (65nm G+)
125M vertices per second (65nm LP) to 188M vertices per second (65nm G+)
250M pixels per second [textured] (65nm LP) to 375M pixels per second [textured] (65nm G+)
250M pixels per second [depth] (65nm LP) to 375M pixels per second [depth] (65nm G+)
** Based on a push-button implementation flow. Higher performance is possible with a custom flow.
TSMC 65nm Characteristics
3.38 square millimeter synthesis area (including memory)
1.5 million standard cells (ND2D1 gates)
60 KB memory (no dual port memories)
92 mW active power at 250MHz (65nm LP)
0.07 mW leakage power at 250MHz (65nm LP)
Features
- 3D graphics API support
- OpenGL ES 2.0
- OpenGL ES 1.1
- Vector graphics API support
- OpenVG 1.1
- 2D graphics API support
- DirectFB 1.4x
- DirectDraw
- GDI
- Adobe Flash acceleration
- Operating systems
- Linux 2.6.21 and higher
- Windows CE 5.0 and higher
- Development platforms
- Android
- Embedded Linux
- Windows Mobile 6.1
- Hardware features
- High performance, small area, low power
- Low bandwidth requirement
- Minimal CPU loading
- Screen resolutions support HD 1080p and beyond
- 4x multisample anti-aliasing (MSAA)
- Ultra-threaded unified shader
- Full GLSL ES shader language support
- IEEE 32-bit floating-point computation supports long shader instructions, exceeds Khronos high precision specification
- 8 programmable elements per primitive
- Bump mapping and up to 24-bit depth map
- Texture support includes 2D, cubic environment, projective, YUY2, and high dynamic range (HDR)
- Depth texturing for high quality shadow generation
- Point sample, bi-linear, and tri-linear texture filtering
- Support for compressed texture standards ETC1 and DXT1-DXT5
- Support for 4 vertex shader and 8 pixel shader simultaneous textures
- 8K x 8K rendering target and texture size
- Depth and two-sided stencil compare
- AHB/AXI or OCP Interface
- Power management features
- Low power CMOS technology
- -> Automatic clock gating
- -> Adaptive frequency and voltage scaling
- Simple programming interface
- Full-Function Programmable 2D Pipeline *
- Line draw, Rectangle fill, Clear, Bit blit, Stretch blit, Filter blit, Monochrome expansion *
- ROP2, ROP3, ROP4 (transparency) support *
- Full alpha blending support
- YUV-to-RGB conversion and 8-bit color lookup *
- 90°,180°, 270° rotation on every primitive *
- Supports up to 255 rectangles per primitive call *
- High-quality image and video scalar (scaler) *
- ->Filter kernel size is programmable from 1x1 to 9x9
- ->Vertical and horizontal kernel sizes are independently programmable
- ->Filter coefficient can be programmed for any type of filter
- ->Supports 32 filter phases
- Programmable format conversion *
- ->Supports 11 source formats
- ->Standard configuration supports 7 destination formats
- * With optional 2D add-on
- Hardened area in silicon using provided push button implementation kit for off-the-shelf Cadence and/or Synopsys tools: 4.7 sq. mm fully routed using 6 metal layers with timing closed at 375 MHz
Benefits
- Very small die area
- Significantly lower power consumption
- Best graphics performance per sq. millimeter in its class
- Fast, unified shader gives programmers maximum flexibility to create stunning visuals.
- Ultra-threaded architecture virtually eliminates graphics pipeline latency
- Multiple screen resolutions up to HD 1080p and beyond
- Optimized for bandwidth
- Minimal CPU loading
Deliverables
- Synthesizable Verilog RTL
- Cadence and Synopsys physical implementation kits
- Portable RTL test bench including basic graphics tests
- Software development kit including test applications
- Drivers: OpenGL ES 2.0/1.1, OpenVG 1.1, DirectFB, GDI
- Documentation
- Optional core and interface customization services
Talk to us
Add your Products