COMP 411 Lecture 2: Lecture 2 notes

• School: UNC
• Department: Computer Science
• Course: COMP 411
• Professor: Singh Montek

Document Summary

Topics: defi(cid:374)i(cid:374)g (cid:862)perfor(cid:373)a(cid:374)(cid:272)e(cid:863, performance measures, how to improve performance, performance pitfalls, examples. Design tradeoffs: performance is rarely the sole factor, cost is important too, energy/power consumption is often critical, frequently used compound metrics, performance/ cost (throughput/$) Lecture 2: performance: performance/power (throughput/watt, work/energy (total work done per joule, for batter powered devices. Execution time: elapsed time/wall clock time, counts everything (disk and memory accesses, i/o, etc. ) Performance: pitfalls of using % -- same example above: by how much is a faster than b, by how much is b slower than a. Instruction count for a program: determined by: program, instruction set architecture (isa, compiler. Lecture 2: performance: a(cid:448)erage (cid:272)(cid:455)(cid:272)les per i(cid:374)stru(cid:272)tio(cid:374) (cid:894)(cid:862)cpi(cid:863)(cid:895, determined by cpu hardware. If different instructions have different cpi average cpi affected by instruction mix. Historically: pdp-11, vax, intel 8086, load/store risc machines mips, sparc, powerpc, minimips, modern cpus, pentium, athlon.