Titlebook: Parallel Programming and Compilers; Constantine D. Polychronopoulos Book 1988 Kluwer Academic Publishers 1988 Natural.compiler.parallel pr

使绝缘

抱狗不敢前

Run-Time Overhead,ncerns methodologies for the best utilization of existing parallel computers. In a given machine-program combination, the limitations (overhead etc.) are known and they cannot be altered. The former issue was analyzed in Chapter 3 from the software point of view. The latter case is the focus of this chapter.

Vasoconstrictor

傻瓜

Static Task Scheduling,tively. Most instances of the spreading problem are NP-Complete [GaJo79]. In this chapter we discuss optimal solutions for some instances of high level spreading, and efficient heuristics for the intractable cases.

abolish

Speedup Bounds for Parallel Programs,uss how one can compute speedup values for inter and intra-task parallelism. Then we consider nested doacross loops and give a generalized speedup formula. Let us state the assumptions and describe the model before we proceed to the derivation of speedup formulas.

Gratuitous

https://doi.org/10.1007/978-1-4613-1077-8Natural; compiler; parallel programming; processor; programming; scheduling; supercomputer