Based On The Transplantation Of Of Tms320dm642 The Avs Decoder Optimized

As next generation video codec standard, AVS has been widely accepted and used in many applications.AVS adopts a number of advanced technologies difference from the previous standards .In addition to imprvod coding efficiency and coding performance, other capabilities of the new standsrd are also enhanced.But improved coding efficiency comes at the cost of higher computational complexity.It is full of challenge to develop an embedded real-time video decoder with the limited on-chip memory space.The TMS320DM642 is the highest-performance fixed-point DSP from Texas Instruments.the DSP core processor has 64 general purpose 32-bit registers and eight hightly independent functional units-two multipliers and six arithmetic logic units include new instructions to accelerate the performance in video and imaging applications and extend the parallelism of the VelociTI architecture.The DM642 uses a two-level cache-based architecture and has a powerful and diverse set peripherals.It is specified that how to migrate AVS testing codes RM50d from PC platform to the wintech DM642 platform and how to optimize the migrated DECODER at length.During the migration,the memory assigmnent,aecess mode and variable definition etc.must be modified in context of the diffrence between the VC compiler of PC and DSP,and the scareity of DSP resouree.Program optimization is the key point in the realization of AVS DECODER. How to aceelerate the decoding speed by making maximum use of advnatages of C64 porcessor is the most important issue in this paper.AVS testing codes pay too mueh attention to readability,ignoring the realtime requirement as a video DECODER.As a result,time--wasted in some modules are quite a little.To speed up decoding time,some measures of embedded system programming and optimization are used in this paper.A series of optimization skills such as the adjustment of the stureture of codes,the modifieation of memory assigmnent,data packing, linear assembly and application of pipeline are adopted in the optimization procedures.Compared the result by the first migrated version of program with the results by the optimized one,the clock cycles of bottleneck functions have significantly decreased.