Texas Instruments SPRAA56 appendix Expected and Measured STS Benchmarks

Page 24

SPRAA56

In the input and output tasks, Cell0 is the color conversion routine. In the processing task, Cell0 is the encoder and Cell1 is the decoder. The expected values for color conversion routines are given as 2-5 ms, typical values for an optimized color conversion routine. Where no expected value was available, the expected value is "—".

Table 1. Expected and Measured STS Benchmarks

STS Benchmark

Expected Value

Measured Value

Your Measurement

 

 

 

 

tskInput

100s of cycles

60,472 instructions

 

 

 

 

 

tskOutput

100s of cycles

11,482,597 instructions

 

 

 

 

 

tskVideoProcess

100s of cycles

24,308 instructions

 

 

 

 

 

tskControl

100s of cycles

702,097 instructions

 

 

 

 

 

stsInVidPeriod

33.33ms

33.26 ms

 

 

 

 

 

stsInVidTotal

1.95 ms

 

 

 

 

 

stsInVidCell0

2-5ms

1.95 ms

 

 

 

 

 

stsInVidWait0

<stsOutVidCell0>

4.75 ms

 

 

 

 

 

stsInVidBusUtil

28,512,000 Bps

 

 

 

 

 

stsOutVidPeriod

33.33ms

33.29 ms

 

 

 

 

 

stsOutVidTotal

2.43 ms

 

 

 

 

 

stsOutVidCell0

2-5ms

2.41 ms

 

 

 

 

 

stsOutVidWait0

<33ms

30.35 ms

 

 

 

 

 

stsOutVidBusUtil

28,512,000 Bps

 

 

 

 

 

stsProcPeriod

33.33ms

33.26 ms

 

 

 

 

 

stsProcTotal

Cell0 + Cell1

24.07 ms

 

 

 

 

 

stsProcCell0

18.97 ms

 

 

 

 

 

stsProcCell1

5.09 ms

 

 

 

 

 

stsProcNframes

1 second (30 frames)

498.84 ms

 

 

 

 

 

stsProcBusUtil

26,926,600 Bps

 

 

 

 

 

The typical expected values for task scheduling latency are on the order of a few hundred cycles, so those benchmarks were gathered in units of instructions rather than milliseconds. Because of the architecture of the video example, where the data tasks all have equal priority, the processing and output task can spend significant time waiting on tasks that are already running. This skews the scheduling latency benchmark higher for all three of the data stream tasks (tskInput, tskOutput, and tskVideoProcess). This can be observed in the Execution Graph by noting the amount of time the tasks remain in the ready state while waiting for currently executing tasks to complete.

24DSP/BIOS Real-Time Analysis (RTA) and Debugging Applied to a Video Application

Image 24
Contents Modifications to the Base Example RTA Techniques for Performance MeasurementViewing Benchmarks in the Instrumented Application References Appendix A. Performance ImpactImportant Benchmarks for Video Applications FiguresBase Application Overview SPRAA56TskInput DSP/BIOS and RF5 Components Used 1 LOG2 STS 3 TRC4 UTL Modifications to the Base Example Requirements for Viewing RTA BenchmarksSplitting the Encode and Decode CELLs Adding the Control TSK and MBX CommunicationQuerying the H.263 Encoder for Status TskInputTskO utput Controlling the Frame Rate RTA Techniques for Performance Measurement Measuring Function Execution Time with the UTL ModuleMeasuring Task Scheduling Latencies Measuring End-to-End LatenciesMeasuring the Frame Rate Programmatic Measurement of Total CPU Load Memory Bus Utilization 720*480 = 345,600 B 86,400 B14,400 B External memoryBitrate and Frame Type Methods for Transmitting Measured Performance Data Viewing Benchmarks in the Instrumented Application Application-Specific Control via GEL Scripts in CCStudioRequirements Running the Application Load the h263loopbackrta.out programSPRAA56 Interpreting the Benchmarks Expected Values for the STS Objects Expected and Measured STS Benchmarks Debug Mode Expected Values Delivered to the Message LogControlling the Run-Time Parameters Dynamically Expected and Measured Logged BenchmarksReferences Capture and Display Task BenchmarkingAppendix A. Performance Impact Overhead of Performance Measurement TechniquesRTA Effects on CPU Load Measured Performance of Benchmarking TechniquesMemory Footprint Memory Footprint DetailsImportant Notice