LINUX Clouds

MPSTAT # mpstat -P ALL 1 01:39:35 AM  CPU    %usr   %nice    %sys %iowait    %irq   %soft  %steal  %guest   %idle 01:39:36 AM  all    0.60    0.00    1.03    0.00    0.00    0.07    0.00    4.57   93.73 01:39:36 AM    0    6.00    0.00    4.00    0.00    0.00    0.00    0.00    5.00   85.00 01:39:36 AM    1    1.98    0.00    2.97    0.00    0.00    0.00    0.00    8.91   86.14 01:39:36 AM    2    3.85    0.00    5.77    0.00    0.00    0.96    0.00    0.00   89.42 01:39:36 AM    3    0.00    0.00    1.55    0.00    0.00    0.00    0.00    3.10   95.35 01:39:36 AM    4    0.96    0.00    1.92    0.00    0.00    0.00    0.00    3.85   93.27 01:39:36 AM    5    0.00    0.00    0.00    0.00    0.00    0.00    0.00    3.03   96.97 01:39:36 AM    6    0.00    0.00    2.00    0.00    0.00    0.00    0.00   19.00   79.00 01:39:36 AM    7    0.98    0.00    0.98    0.00    0.00    0.00    0.00    4.90   93.14 01:39:36 AM    8    0.00    0.00    0.00    0.00    0.00    0.00    0.

top, cont. • 1. top can miss: • short-lived processes • kernel threads (tasks), unless included (see top options) • 2. analyzing high CPU processes: • identify why – profile code path • identify what – execution or stall cycles • High %CPU time may be stall cycles on memory I/O – upgrading to faster CPUs doesn’t help! Sunday, February

Cache Performance Last time: intro, idea of CPU time for a program. Cache miss is “lost time” to the system, counted officially as “CPU time” since it’s handled completely by the CPU. The CPU just rates as slower because of all the cache misses. We see it’s all up to the CPU to manage the bus and access to main memory. Here’s the breakdown: °          CPU time = (CPU exec clock cycles + memory stall cycles) * clock cycle time °          Memory stalls = read stalls + write stalls °          CPU time = (CPU exec clock cycles + memory stall cycles) * clock cycle time °          Memory stalls = read stalls + write stalls °          Read stall cycles = reads per program * read miss rate * read miss penalty °          Write stalls for write-through •           write stall cycles = Writes/program * write miss rate * write miss penalty + write buffer stalls •           Write stalls for write-back: more complicated, defer °          Write buffer stalls are usually small

take "cost" as a variable [root@noisecover-dr Scripts]# cost=0 [root@noisecover-dr Scripts]# printf $cost --> print cost value 0[root@noisecover-dr Scripts]# --> prompt will come on same line