diff options
Diffstat (limited to 'doc/src/platforms/emb-hardwareacceleration.qdocinc')
-rw-r--r-- | doc/src/platforms/emb-hardwareacceleration.qdocinc | 129 |
1 files changed, 129 insertions, 0 deletions
diff --git a/doc/src/platforms/emb-hardwareacceleration.qdocinc b/doc/src/platforms/emb-hardwareacceleration.qdocinc new file mode 100644 index 0000000..3851628 --- /dev/null +++ b/doc/src/platforms/emb-hardwareacceleration.qdocinc @@ -0,0 +1,129 @@ + +\section1 Hardware Acceleration + +When designing applications for embedded devices the choice often stands +between graphics effects and performance. On most devices, you cannot have +both simply because the hardware needed for such operations just is not +there. Still a growing number of devices use hardware dedicated to graphics +operations to improve performance. + +Using graphics acceleration hardware is more power efficient than using the +CPU. The reason for this is that the CPU might require a clock speed that +is up to 20 times higher than the GPU, achieving the same results. E.g. a +typical hardware accelerated mobile graphics unit can rasterize one or two +bilinear texture fetches in one cycle, while a software implementation +takes easily more than 20 cycles. Graphics hardware generally have a much +lower clock speed and memory bandwidth and different level of acceleration +than desktop GPUs. One example is that many GPUs leave out transformation +and lighting from the graphics pipeline and only implements rasterization. + +So the key to write good applications for devices is therefore to limit the +wow factor down to what the target hardware can handle, and to take +advantage of any graphics dedicated hardware. Qt provides several ways to +both render advanced effects on the screen and speed up your application +using hardware accelerated graphics. + +\tableofcontents + +\section2 Qt for Embedded Graphics pipeline + +Qt uses QPainter for all graphics operations. By using the same API +regardless of platform, the code can be reused on different devices. +QPainter use different paint engines implemented in the QPaintEngine API to +do the actual painting. + +The QPaintEngine API provides paint engines for each window system and +painting framework supported by Qt. In regards to Qt for Embedded, this +also includes implementations for OpenGL ES versions 1.1 and 2.0, as well +as OpenVG and DirectFB(Embedded Linux only). + +By using one of these paint engines, you will be able to improve the +graphics performance of your Qt application. However, if the graphics +operations used are not supported, this might as well be a trap, slowing +down your application significantly. This all depends on what kind of +graphics operations that are supported by the target devices hardware +configuration. + +\image platformHWAcc.png + +The paint engine will direct all graphics operations supported by the +devices hardware to the GPU, and from there they are sent to the +framebuffer. Unsupported graphics operations falls back to the +QRasterPaintEngine and are handled by the CPU before sent to the +framebuffer. In the end, the operating system sends the paint updates off +to the screen/display. The fallback operation is quite expensive in regards +to memory consumption, and should be avoided. + +\section2 Hardware configuration requirements + +Before implementing any application using hardware acceleration, it is wise +to get an overview of what kind of hardware accelerated graphics operations +that are available for the target device. + +\note On devices with no hardware acceleration, Qt will use +QRasterPaintEngine, which handles the acceleration using software. On +devices supporting OpenGL ES, OpenVG or DirectFB(not supported by Windows +CE), Qt will use the +respective paint engines to accelerate painting. However, hardware +configurations that only support a limited set of hardware acceleration +features, might slow the application graphics down rather than speeding it +up when using unsupported operations that must fall back to the raster +engine. + +\section3 Different architectures + +Based on the architecture used in a device we can make a recommendation on +which hardware acceleration techniques to use. There are mainly two +different architectures on embedded devices. These are devices with a +Unified Memory Architecture (UMA), and devices with dedicated graphics +memory. Generally, high-end devices will have dedicated graphics memory. +Low-end devices will just use system memory, sometimes reserving a memory +region and sometimes not. + +In addition to this, we can categorize the devices into five types based on +the different graphics operations supported by their hardware. + +\list 1 + \o No support for graphics acceleration. + \o Support for blitter and alpha blending. + \o Support for path based 2D vector graphics. + \o Support for fixed function 3D graphics. + \o Support for programmable 3D graphics. +\endlist + +Based on these characteristics the table below recommends which paint +engines to use with the different types of hardware configurations. + +\section3 Recommended use of hardware acceleration based on hardware + + \table + \header + \o Type + \o UMA + \o Non-UMA + \row + \o \bold {None} + \o Qt Raster Engine + \o Qt Raster Engine + \row + \o \bold {Blitter} + \o DirectFB + \o DirectFB + \row + \o \bold {2D Vector} + \o OpenVG + \o OpenVG + \row + \o \bold {Fixed 3D} + \o OpenGL (ES) 1.x + \o OpenGL (ES) 1.x + \row + \o \bold {Programmable 3D} + \o OpenGL (ES) 2.x + \o OpenGL (ES) 2.x + + \endtable + +\note Since the DirectFB API is quite primitive, the raster paint engine +handles most of the operations. +\note Blitter and Alpha blending is currently not supported on Windows CE. |