This Seminar Report, describes an implementation of Vectorizing Compiler for Intelís MMX ( MultiMedia Extensions).
This compiler would identify data parallel sections of the code using scalar expansion and array dependence analysis. To enhance the scope for application of the subword semantics, this compiler performs several code transformations.

These include strip mining, scalar expansion, grouping and reduction, loop fission and distribution.There after inline assembly
instructions corresponding to the data parallel sections are generated. This compiler uses Stanford University Intermediate
Format(SUIF),a public domain compiler tool, for implementation.



The performance of the code generated by this compiler is evaluated for a multimedia benchmarks. Initial performance results
reveal that, this compiler generated code produces a reasonable performance improvement (speedup of 2 to6.5) over the
code generated without the vectorizing transformations/inline assembly.

To know more on Vectorizing Compiler For Multimedia Extension download the report attached below
.