Multi-Core Software

Inside the Intel® 10.1 Compilers: New Threadizer and New Vectorizer for Intel® Core™2 Processors

CONCLUSION

The Intel® 10.1 C++/Fortran compiler features various advanced compiler optimizations to leverage the enhanced capabilities of Intel® Core™2 Duo and Quad processors. Threadization exploits thread-level parallelism in serial programs; vectorization exploits SIMD-based vector-level parallelism; and advanced code generation exploits important micro-architectural features for gaining a higher performance. This paper presented the implementation of the new threadizer and vectorizer and an overview of advanced code generation that specifically leverages the Intel® Core™ micro-architecture.

Performance validation was conducted with a large set of real-life industry-standard benchmarks. It was shown that advanced optimizations of the Intel C++/Fortran compiler can obtain further improvements over optimized code, with contributions from threadization, vectorization, loop optimizations, and target-specific code generation. Furthermore, these optimizations were added in a manner that still allows for our overall goal of continuing to generate code that runs well across all processors.

More information on Intel® high-performance compilers for Intel® Architectures can be found at the Intel website http://www.intel.com/software/products/.

Back to Top