FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2275205953> ?p ?o ?g. }

• W2275205953 abstract "Multimedia extensions are nearly ubiquitous in today's general-purpose processors. These extensions consist primarily of a set of short-vector instructions that apply the same opcode to a vector of operands. This design introduces a data-parallel component to processors that exploit instruction-level parallelism, and presents an opportunity for increased performance. In fact, ignoring a processor's vector opcodes can leave a significant portion of the available resources unused. In order for software developers to find short-vector instructions generally useful, the compiler must target these extensions with complete transparency and consistent performance. This thesis develops compiler techniques to target short-vector instructions automatically and efficiently. One important aspect of compilation is the effective management of memory alignment. As with scalar loads and stores, vector references are typically more efficient when accessing aligned regions. In many cases, the compiler can glean no alignment information and must emit conservative code sequences. In response, I introduce a range of compiler techniques for detecting and enforcing aligned references. In my benchmark suite, the most practical method ensures alignment for roughly 75% of dynamic memory references. This thesis also introduces selective vectorization, a technique for balancing computation across a processor's scalar and vector resources. Current approaches for targeting short-vector instructions directly adopt vectorizing technology first developed for supercomputers. Traditional vectorization, however, can lead to a performance degradation since it fails to account for a processor's scalar execution resources. I formulate selective vectorization in the context of software pipelining. My approach creates software pipelines with shorter initiation intervals, and therefore, higher performance. In contrast to conventional methods, selective vectorization operates on a low-level intermediate representation. This technique allows the algorithm to accurately measure the performance trade-offs of code selection alternatives. A key aspect of selective vectorization is its ability to manage communication of operands between vector and scalar instructions. Even when operand transfer is expensive, the technique is sufficiently sophisticated to achieve significant performance gains. I evaluate selective vectorization on a set of SPEC FP benchmarks. On a realistic VLIW processor model, the approach achieves whole-program speedups of up to 1.35× over existing approaches. For individual loops, it provides speedups of up to 1.75×. (Copies available exclusively from MIT Libraries, Rm. 14-0551, Cambridge, MA 02139-4307. Ph. 617-253-5668; Fax 617-253-1690.)" @default.
• W2275205953 created "2016-06-24" @default.
• W2275205953 creator A5046791216 @default.
• W2275205953 creator A5049679335 @default.
• W2275205953 date "2006-01-01" @default.
• W2275205953 modified "2023-09-27" @default.
• W2275205953 title "Compilation techniques for short-vector instructions" @default.
• W2275205953 cites W127897580 @default.
• W2275205953 cites W1480457706 @default.
• W2275205953 cites W1491178396 @default.
• W2275205953 cites W1494930385 @default.
• W2275205953 cites W1495550651 @default.
• W2275205953 cites W1555915743 @default.
• W2275205953 cites W1560132800 @default.
• W2275205953 cites W1591319746 @default.
• W2275205953 cites W1963718362 @default.
• W2275205953 cites W1965638741 @default.
• W2275205953 cites W1966324811 @default.
• W2275205953 cites W1966691503 @default.
• W2275205953 cites W1974200321 @default.
• W2275205953 cites W1976158744 @default.
• W2275205953 cites W1979072566 @default.
• W2275205953 cites W1984972320 @default.
• W2275205953 cites W1986535055 @default.
• W2275205953 cites W1992908040 @default.
• W2275205953 cites W1998425871 @default.
• W2275205953 cites W2001984859 @default.
• W2275205953 cites W2003368909 @default.
• W2275205953 cites W2005128309 @default.
• W2275205953 cites W2005343162 @default.
• W2275205953 cites W2010418442 @default.
• W2275205953 cites W2012512596 @default.
• W2275205953 cites W2022711417 @default.
• W2275205953 cites W2024804757 @default.
• W2275205953 cites W2034579607 @default.
• W2275205953 cites W2038198320 @default.
• W2275205953 cites W2066951238 @default.
• W2275205953 cites W2069703635 @default.
• W2275205953 cites W2079834329 @default.
• W2275205953 cites W2082265286 @default.
• W2275205953 cites W2087256755 @default.
• W2275205953 cites W2095117703 @default.
• W2275205953 cites W2099404643 @default.
• W2275205953 cites W2100914138 @default.
• W2275205953 cites W2107139334 @default.
• W2275205953 cites W2111394443 @default.
• W2275205953 cites W2112482891 @default.
• W2275205953 cites W2114067856 @default.
• W2275205953 cites W2115184416 @default.
• W2275205953 cites W2115567860 @default.
• W2275205953 cites W2118382442 @default.
• W2275205953 cites W2121398690 @default.
• W2275205953 cites W2123412205 @default.
• W2275205953 cites W2130983588 @default.
• W2275205953 cites W2132929529 @default.
• W2275205953 cites W2137249790 @default.
• W2275205953 cites W2137857636 @default.
• W2275205953 cites W2139802090 @default.
• W2275205953 cites W2140311411 @default.
• W2275205953 cites W2140515042 @default.
• W2275205953 cites W2140656851 @default.
• W2275205953 cites W2143659699 @default.
• W2275205953 cites W2144481293 @default.
• W2275205953 cites W2144643978 @default.
• W2275205953 cites W2147423491 @default.
• W2275205953 cites W2148492728 @default.
• W2275205953 cites W2152687294 @default.
• W2275205953 cites W2154882361 @default.
• W2275205953 cites W2155583896 @default.
• W2275205953 cites W2156413870 @default.
• W2275205953 cites W2158967724 @default.
• W2275205953 cites W2161455936 @default.
• W2275205953 cites W2164215436 @default.
• W2275205953 cites W2165964351 @default.
• W2275205953 cites W2166155386 @default.
• W2275205953 cites W2168922271 @default.
• W2275205953 cites W2576288252 @default.
• W2275205953 cites W27956500 @default.
• W2275205953 cites W2913795363 @default.
• W2275205953 cites W37253703 @default.
• W2275205953 cites W406491355 @default.
• W2275205953 cites W63944998 @default.
• W2275205953 cites W204452631 @default.
• W2275205953 cites W2132886713 @default.
• W2275205953 cites W2498227530 @default.
• W2275205953 hasPublicationYear "2006" @default.
• W2275205953 type Work @default.
• W2275205953 sameAs 2275205953 @default.
• W2275205953 citedByCount "1" @default.
• W2275205953 crossrefType "dissertation" @default.
• W2275205953 hasAuthorship W2275205953A5046791216 @default.
• W2275205953 hasAuthorship W2275205953A5049679335 @default.
• W2275205953 hasConcept C111919701 @default.
• W2275205953 hasConcept C140763907 @default.
• W2275205953 hasConcept C150552126 @default.
• W2275205953 hasConcept C161824985 @default.
• W2275205953 hasConcept C165696696 @default.
• W2275205953 hasConcept C169590947 @default.
• W2275205953 hasConcept C170595534 @default.
• W2275205953 hasConcept C173608175 @default.

• next