Sign in to follow this  
ITMan

FIO Benchmark

Recommended Posts

Hi friends,

I am testing with FIO, I need to help for interpret outputs, for example I have to output FIO,
1. iodepth=1
2. iodepth=16

I have two question:
1. for different iodepths, I get same IOPS , in both output IOPS is 75, 
why both have a same IOPS?

my commands is:

# fio --filename=/dev/sdb --direct=1 --rw=randrw --ioengine=libaio --bs=4K  --name=4K_rw --time_based  --runtime=180   –ramp_time=30 –invalidate=1  --iodepth=1   --rwmixwrite=100 –fsync=1[/code]

output:

test4K: (g=0): rw=randrw, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T) 4096B-4096B, ioengine=libaio, iodepth=1
fio-3.7
Starting 1 process
	test4K: (groupid=0, jobs=1): err= 0: pid=3115: Sat Jun 23 02:09:28 2018
  write: [b]IOPS=75[/b], BW=304KiB/s (311kB/s)(53.4MiB/180012msec)
    slat (usec): min=12, max=454, avg=24.20, [b]stdev= 4.65[/b]
    clat (usec): min=7, max=397, avg=243.12, stdev=18.27
     lat (usec): min=211, max=466, avg=267.94, stdev=18.39
    clat percentiles (usec):
     |  1.00th=[  219],  5.00th=[  225], 10.00th=[  227], 20.00th=[  231],
     | 30.00th=[  235], 40.00th=[  239], 50.00th=[  241], 60.00th=[  245],
     | 70.00th=[  247], 80.00th=[  251], 90.00th=[  258], 95.00th=[  262],
     | 99.00th=[  334], 99.50th=[  351], 99.90th=[  375], 99.95th=[  383],
     | 99.99th=[  396]
   bw (  KiB/s): min=  272, max=  344, per=100.00%, avg=303.66, stdev=14.87, samples=360
   iops        : min=   68, max=   86, avg=75.91, stdev= 3.72, samples=360
  lat (usec)   : 10=0.01%, 250=76.67%, 500=23.32%
  fsync/fdatasync/sync_file_range:
    sync (nsec): min=377, max=22007, avg=657.98, stdev=366.05
    sync percentiles (nsec):
     |  1.00th=[  474],  5.00th=[  644], 10.00th=[  644], 20.00th=[  644],
     | 30.00th=[  644], 40.00th=[  644], 50.00th=[  652], 60.00th=[  652],
     | 70.00th=[  652], 80.00th=[  660], 90.00th=[  676], 95.00th=[  676],
     | 99.00th=[  684], 99.50th=[  748], 99.90th=[  780], 99.95th=[  828],
     | 99.99th=[20608]
  cpu          : usr=0.22%, sys=0.49%, ctx=27330, majf=0, minf=759
  IO depths    : 1=233.3%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwts: total=0,13665,0,15939 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=1
	Run status group 0 (all jobs):
  WRITE: bw=304KiB/s (311kB/s), 304KiB/s-304KiB/s (311kB/s-311kB/s), io=53.4MiB (55.0MB), run=180012-180012msec
	Disk stats (read/write):
  sdb: ios=92/31873, merge=0/0, ticks=9/208643, in_queue=208637, util=99.29%
  
 ***************************************************************************************************
 ***************************************************************************************************
  second command:

  #fio --filename=/dev/sdb --direct=1 --rw=randrw --ioengine=libaio --bs=4K  --name=4K_rw --time_based  --runtime=180   –ramp_time=30 –invalidate=1  --iodepth=16  --rwmixwrite=100 –fsync=1


  output :
 test4K: (g=0): rw=randrw, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T) 4096B-4096B, ioengine=libaio, iodepth=16
fio-3.7
Starting 1 process

test4K: (groupid=0, jobs=1): err= 0: pid=3168: Sat Jun 23 02:22:46 2018
  write: IOPS=75, BW=304KiB/s (311kB/s)(53.5MiB/180015msec)
    slat (nsec): min=9892, max=71305, avg=24227.85, stdev=3285.42
    clat (msec): min=17, max=273, avg=197.41, stdev=15.72
     lat (msec): min=17, max=273, avg=197.43, stdev=15.72
    clat percentiles (msec):
     |  1.00th=[  163],  5.00th=[  174], 10.00th=[  180], 20.00th=[  186],
     | 30.00th=[  190], 40.00th=[  194], 50.00th=[  197], 60.00th=[  201],
     | 70.00th=[  205], 80.00th=[  211], 90.00th=[  218], 95.00th=[  224],
     | 99.00th=[  239], 99.50th=[  243], 99.90th=[  259], 99.95th=[  266],
     | 99.99th=[  271]
   bw (  KiB/s): min=  256, max=  352, per=99.95%, avg=303.83, stdev=14.89, samples=360
   iops        : min=   64, max=   88, avg=75.95, stdev= 3.72, samples=360
  lat (msec)   : 20=0.01%, 50=0.01%, 100=0.04%, 250=99.78%, 500=0.27%
  fsync/fdatasync/sync_file_range:
    sync (nsec): min=227, max=25595, avg=659.43, stdev=531.37
    sync percentiles (nsec):
     |  1.00th=[  612],  5.00th=[  620], 10.00th=[  628], 20.00th=[  628],
     | 30.00th=[  628], 40.00th=[  636], 50.00th=[  644], 60.00th=[  652],
     | 70.00th=[  660], 80.00th=[  660], 90.00th=[  668], 95.00th=[  668],
     | 99.00th=[  676], 99.50th=[  676], 99.90th=[  780], 99.95th=[17536],
     | 99.99th=[24704]
  cpu          : usr=0.15%, sys=0.33%, ctx=13676, majf=0, minf=310
  IO depths    : 1=0.1%, 2=0.1%, 4=0.1%, 8=0.1%, 16=232.9%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.1%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwts: total=0,13673,0,15921 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=16

Run status group 0 (all jobs):
  WRITE: bw=304KiB/s (311kB/s), 304KiB/s-304KiB/s (311kB/s-311kB/s), io=53.5MiB (56.1MB), run=180015-180015msec

Disk stats (read/write):
  sdb: ios=92/31847, merge=0/0, ticks=15/212950, in_queue=212953, util=99.53%


  
  Thanks in Advance,

Share this post


Link to post
Share on other sites

What are you testing? Also I've never used the -invalidate flag before, what is that used with?

  • Thanks 1

Share this post


Link to post
Share on other sites

Hi Kevin,

Thanks,

I use randwrite for random write with 100% write  and

--invalidate flag can be ignore,

 

Share this post


Link to post
Share on other sites

Context of what your are testing will help greatly on this. 

  • Thanks 1

Share this post


Link to post
Share on other sites

Hi Admin

ok, My quesion is  :

which measure is correct for latency in below output:

slat (usec): min=12, max=454, avg=24.20, [b]stdev= 4.65[/b]
    clat (usec): min=7, max=397, avg=243.12, stdev=18.27
     lat (usec): min=211, max=466, avg=267.94, stdev=18.39
is really latency is microeconomics?
Thanks,

Share this post


Link to post
Share on other sites

This is the line for full latency, and correct that it is in microseconds:

 

     lat (usec): min=211, max=466, avg=267.94, stdev=18.39
  • Thanks 1

Share this post


Link to post
Share on other sites

Hi Kevin,

Thanks,

as last question,

I post in first topic , i  for different iodepths, I get same IOPS , in both output IOPS is 75, 

why for different  iodepths, it get same IOPS?

 

Thanks in advance,

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this