Search
Skip to Search Results-
2006
Amaral, Nelson, Niewiadomski, Robert
Technical report TR06-06. Research on the performance of implicit k-heaps has shown that aligning data with cache lines and increasing heap arity are effective techniques for improving the data reference locality of heap operations. The technique of tree blocking has long been used to enhance the...
-
2006
Amaral, Nelson, Touati, Sid-Ahmed-Ali, Berube, Paul, Huynh, Johnny
Technical report TR06-04. In digital signal processors (DSPs) variables are accessed using k address registers. The problem of finding a memory layout, for a set of variables, that minimizes the address-computation overhead is known as the General Offset Assignment (GOA) Problem. The most common...
-
2004
Technical report TR04-26. Switch-case statements provide a concise way to express multi-way branching control flow semantics. Switch-cases are common in programs, including script parsers, compilers and virtual machines. Because they occur in widely used programs, optimization for switch-cases...
-
2004
Technical report TR04-18. Large functions appear frequently in today's applications. They present challenges to modern compilers not only because they increase the compilation cost including time and resources but also they usually degrade the quality of the final executables. To reduce these...
-
Using SIMD Registers and Instructions to Enable Instruction-Level Parallelism in Sorting Algorithms
Download2007
Amaral, Nelson, Niewiadomski, Robert, Furtak, Timothy
Technical report TR07-02. Most contemporary processors offer some version of Single Instruction Multiple Data (SIMD) machinery -- vector registers and instructions to manipulate data stored in such registers. The central idea of this paper is to use these SIMD resources to improve the...