To improve energy efficiency, PANTA family components might be used. Figure 2 shows an example of a graphics subsystem architecture which utilizes two PANTA DP30 display processors equipped with the PANTA CP20 scaling coprocessors. In this case, the GPU is responsible only for rendering graphics layers pixels and sending them to the frame buffer. The GPU is therefore significantly offloaded because YUV to RGB conversion, alpha-blending, scaling and rotation are handled by the PANTA DP30 and PANTA CP20 units.
Figure 2: Multi-display subsystem with PANTA components
In the presented system, the video layer converted by the PANTA DP30 to the RGB format is composed with the other graphics layers and directly displayed on the external panel. At the same time the composed frame is down-scaled from 1080p to 720p by the PANTA CP20 module and returned to the frame buffer. The second PANTA display processor fetches the scaled frame and rotates it by 90 degrees before transmitting it to the local display. Thanks to the PANTA IP scaling and rotation capabilities these additional two operations are not executed in the GPU, and therefore overall power dissipation in the graphics subsystem is significantly decreased. Additionally, since the size of graphics data stored in the frame buffer depends on frame resolution and format, usage of the PANTA components allows reduction of system memory bandwidth up to 40 percent against a typical multi-display solution shown in Figure 1
. Total power consumption of PANTA components presented in this use case and implemented in the 40nm LP process is less than 30 mW.
Going further, PANTA display processors enable even higher reduction of power consumption. In some cases all the video and graphics processing tasks can be executed by the PANTA components, thus allowing complete switching off the GPU. Figure 3
presents a use case where the PANTA IP displays decoded video forwarded to the frame buffer by a video decoder. Such a stream requires only rotation and YUV to RGB conversion, and thus eliminates the need to engage the GPU. In this case, the power consumption of the PANTA DP30 component does not exceed 6 mW for the implementation in the 40nm LP process.
Figure 3: PANTA DP-aided display subsystem without GPU
Click on image to enlarge
Because of 3D gaming and HD video playback becoming essential features in mobile devices, it’s crucial that the described video pipeline is power-optimized at the SoC level. Users expect smartphones and tablets to provide decent battery life, also during multimedia activities.
Evatronix PANTA processors improve the power efficiency of the display subsystem without any compromise for the user. It is achieved thanks to the IP performing display specific tasks instead of the GPU and minimizing memory bandwidth of the video/graphics subsystem.
To find out more about the Evatronix PANTA DP display processors, visit Evatronix website at www.evatronix.com/ip
. Selecting the right battery
. Implementing battery charger using Li-ion
. The case for automated battery management
. Energy harvesting & battery pack protection from shunt charger IC systems
. Unique input regulation loop simplifies solar panel battery charging
. Battery Charge Control: Dedicated ICs vs Microcontrollers
. Shunt charger ICs harvest energy, protect batteries
. Battery Management Systems for Large Lithium-Ion Battery Packs (Part 1 of 3)
. Battery Management Systems for Large Lithium-Ion Battery Packs (Part 2 of 3)
. Battery Management Systems for Large Lithium-Ion Battery Packs (Part 3 of 3)