1260442079 - Popular Views of Performance PC Magazine Oct...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon
Computer Organization CDA 3103 Dr. Hassan Foroosh Dept. of Computer Science UCF © Copyright Hassan Foroosh 2004 Popular Views of Performance ± PC Magazine Oct 7, 1997 ± pp. 22-23 ad by SAG electronics: “SAG offers the best price/performance” attributed to Network Computing June 1, 1997 ± P. 48 Benchmark Tests: “These Pentium MMX/233 notebooks were great performers, posting Business Winstone scores that averaged about 21% higher than the average scores of 166 MHz notebooks in our August roundup” ± p. 216 ad by DEC: “…you’ll grow to love running Windows NT ® software on the world’s fastest microprocessor” ± What do you think these statements mean? ± How would you determine whether they are true? Today’s Lecture: Performance ± Performance: ± What is it? ± Why is it important? ± How do we measure it? ± The CPU performance equation ± Execution time as the measure ± What affects execution time ± Examples ± Popular alternative metrics ± Reasons they don’t work ± Choosing good benchmarks ± Choosing bad benchmarks ± Comparing and summarizing performance ± Good and bad practices ± Amdahl's Law Performance is Time to Complete Work Plane Boeing 747 BAC/Sud Concorde Speed 610 MPH 1350 MPH DC to Paris 6.5 hours 3 hours Passengers 470 132 Throughput (PMPH) 286,700 178,200 ± Time to do the task (Execution Time) ± execution time, response time, latency ± Tasks per unit time (sec, minute, . ..) ± throughput, bandwidth
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Performance as Response Time ± Performance is most often measured as response time or execution time for some task. ± “X is n times faster than Y” means Performance(X) Execution Time(Y) –––––––––––––– = –––––––––––––––– = Performance(Y) Execution Time(X) ± Example ± Execution time of program P ± X is 5 sec ± Y is 10 sec. ± X is 2 times faster than Y What Time to Measure? ± Elapsed time, wall-clock time ± Actual time from start to completion ± Depends on CPU, system, I/O, etc. ± Often used in real benchmarks ± Only suitable choice when I/O is included ± CPU Time ± Measure/analyze CPU performance only ± May be suitable when machine is timeshared ± Possibly both user and system component ± User CPU time is our focus for first part of course ± Elapsed time = CPU time + Idle time ± When time is accurately accounted for Why is Performance Important? ± Performance is a key characteristic to evaluate computer systems vs. requirements and competition ± Computers are purchased to complete work ± The computer’s work is valuable to a user’s productivity, education, entertainment, . .. ± A user generally has minimum performance requirements, and she will perceive benefit from better performance ± But it is not the only important characteristic ± Compatibility with software, hardware, data formats, standards ± Cost, size, power ± Reliability, Availability, Serviceability (RAS) ± Extensibility, Security, . ..
Background image of page 2
Image of page 3
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 08/22/2010 for the course CDA 3101 taught by Professor Staff during the Fall '07 term at University of Central Florida.

Page1 / 10

1260442079 - Popular Views of Performance PC Magazine Oct...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online