Page 1Time MeasurementFeb 17, 2004TopicsTopics Time scales Interval counting Cycle counters K-best measurement schemeclass11.ppt15-213“The course that gives CMU its Zip!”– 2 –15-213, s’04Computer Time ScalesTwo Fundamental Time ScalesTwo Fundamental Time Scales Processor: ~10–9sec. External events: ~10–2sec. Keyboard input Disk seek Screen refreshImplicationImplication Can execute many instructions while waiting for external event to occur Can alternate among processes without anyone noticingTime Scale (1 Ghz Machine)1.E-09 1.E-06 1.E-03 1.E+00Time (seconds)1 ns 1 µs 1 ms 1 sInteger AddFP MultiplyFP DivideKeystrokeInterruptHandlerDisk AccessScreen RefreshKeystrokeMicroscopicMacroscopic– 3 –15-213, s’04Measurement ChallengeHow Much Time Does Program X Require?How Much Time Does Program X Require? CPU time How many total seconds are used when executing X? Measure used for most applications Small dependence on other system activities Actual (“Wall”) Time How many seconds elapse between the start and the completion of X? Depends on system load, I/O times, etc.Confounding FactorsConfounding Factors How does time get measured? Many processes share computing resources Transient effects when switching from one process to another Suddenly, the effects of alternating among processes become noticeable– 4 –15-213, s’04“Time” on a Computer Systemreal (wall clock) time= user time (time executing instructions in the user process)+ = real (wall clock) timeWe will use the word “time” to refer to user time.= system time (time executing instructions in kernel on behalf of user process)+= some other user’s time (time executing instructions in different user’s process)cumulative user timePage 2– 5 –15-213, s’04Activity Periods: Light LoadMost of the time spent executing one process Periodic interrupts every 10ms Interval timer Keep system from executing one process to exclusion of others Other interrupts Due to I/O activity Inactivity periods System time spent processing interrupts ~250,000 clock cycles Activity Periods, Load = 10 10 20 30 40 50 60 70 801Time (ms)ActiveInactive– 6 –15-213, s’04Activity Periods: Heavy LoadSharing processor with one other active process From perspective of this process, system appears to be “inactive” for ~50% of the time Other process is executingActivity Periods, Load = 20 10 20 30 40 50 60 70 801Time (ms)ActiveInactive– 7 –15-213, s’04Interval CountingOS Measures Runtimes Using Interval TimerOS Measures Runtimes Using Interval Timer Maintain 2 counts per process User time System time Each time get timer interrupt, increment counter for executing process User time if running in user mode System time if running in kernel mode– 8 –15-213, s’04Interval Counting ExampleAu Au Au As Bu Bs Bu Bu Bu Bu As Au Au Au Au Au Bs Bu Bu Bs Au Au Au As AsA 110u + 40sB 70u + 30s(a) Interval TimingsB BAA A(b) Actual TimesBAABA 120.0u + 33.3sB 73.3u + 23.3s0 10 20 30 40 50 60 70 80 90 100110120130140150160AAu Au Au As Bu Bs Bu Bu Bu Bu As Au Au Au Au Au Bs Bu Bu Bs Au Au Au As AsA 110u + 40sB 70u + 30s(a) Interval TimingsB BAA AAu Au Au As Bu Bs Bu Bu Bu Bu As Au Au Au Au Au Bs Bu Bu Bs Au Au Au As AsA 110u + 40sB 70u + 30s(a) Interval TimingsB BAA A(b) Actual TimesBAABA 120.0u + 33.3sB 73.3u + 23.3s0 10 20 30 40 50 60 70 80 90 100110120130140150160A(b) Actual TimesBAABA 120.0u + 33.3sB 73.3u + 23.3s0 10 20 30 40 50 60 70 80 90 100110120130140150160APage 3– 9 –15-213, s’04Unix time Command0.82 seconds user time 82 timer intervals 0.30 seconds system time 30 timer intervals 1.32 seconds wall time 84.8% of total was used running these processes (.82+0.3)/1.32 = .848time make oseventgcc -O2 -Wall -g -march=i486 -c clock.cgcc -O2 -Wall -g -march=i486 -c options.cgcc -O2 -Wall -g -march=i486 -c load.cgcc -O2 -Wall -g -march=i486 -o osevent osevent.c . . .0.820u 0.300s 0:01.32 84.8% 0+0k 0+0io 4049pf+0w– 10 –15-213, s’04Accuracy of Interval CountingWorst Case AnalysisWorst Case Analysis Timer Interval = δδδδ Single process segment measurement can be off by ±δ±δ±δ±δ No bound on error for multiple segments Could consistently underestimate, or consistently overestimate0 10 20 30 40 50 60 70 80AAMinimumMaximum0 10 20 30 40 50 60 70 80AAMinimumMaximum• Computed time = 70ms• Min Actual = 60 + εεεε• Max Actual = 80 – εεεε– 11 –15-213, s’04Accuracy of Interval Couting (cont.)Average Case AnalysisAverage Case Analysis Over/underestimates tend to balance out As long as total run time is sufficiently large Min run time ~1 second 100 timer intervals Consistently miss 4% overhead due to timer interrupts0 10 20 30 40 50 60 70 80AAMinimumMaximum0 10 20 30 40 50 60 70 80AAMinimumMaximum• Computed time = 70ms• Min Actual = 60 + εεεε• Max Actual = 80 – εεεε– 12 –15-213, s’04Cycle CountersMost modern systems have built in registers that are incremented every clock cycle Very fine grained Maintained as part of process state» In Linux, counts elapsed global time Special assembly code instruction to access On (recent model) Intel machines: 64 bit counter. RDTSC instruction sets %edx to high order 32-bits, %eaxto low order 32-bitsPage 4– 13 –15-213, s’04Cycle Counter PeriodWrap Around Times for 550 MHz machineWrap Around Times for 550 MHz machine Low order 32 bits wrap around every 232/ (550 * 106) = 7.8 seconds High order 64 bits wrap around every 264/ (550 * 106) = 33539534679 seconds 1065 yearsFor 2 GHz machineFor 2 GHz machine Low order 32-bits every 2.1 seconds High order 64 bits every 293 years– 14 –15-213, s’04Measuring with Cycle CounterIdeaIdea Get current value of cycle counter store as pair of unsigned’s cyc_hi and cyc_lo Compute something Get new value of cycle counter Perform double precision subtraction to get elapsed cycles/* Keep track of most recent reading of cycle counter */static unsigned cyc_hi = 0;static unsigned cyc_lo = 0;void start_counter(){/* Get current value of cycle counter */access_counter(&cyc_hi, &cyc_lo);}– 15 –15-213, s’04Accessing the Cycle Cntr.GCC allows inline assembly code with mechanism for matching registers with program variables Code only works on x86 machine compiling with
View Full Document