8 5. Detailed list of parameters
13 1.0 Overview and history
14 ------------------------
15 fio was originally written to save me the hassle of writing special test
16 case programs when I wanted to test a specific workload, either for
17 performance reasons or to find/reproduce a bug. The process of writing
18 such a test app can be tiresome, especially if you have to do it often.
19 Hence I needed a tool that would be able to simulate a given io workload
20 without resorting to writing a tailored test case again and again.
22 A test work load is difficult to define, though. There can be any number
23 of processes or threads involved, and they can each be using their own
24 way of generating io. You could have someone dirtying large amounts of
25 memory in an memory mapped file, or maybe several threads issuing
26 reads using asynchronous io. fio needed to be flexible enough to
27 simulate both of these cases, and many more.
31 The first step in getting fio to simulate a desired io workload, is
32 writing a job file describing that specific setup. A job file may contain
33 any number of threads and/or files - the typical contents of the job file
34 is a global section defining shared parameters, and one or more job
35 sections describing the jobs involved. When run, fio parses this file
36 and sets everything up as described. If we break down a job from top to
37 bottom, it contains the following basic parameters:
39 IO type Defines the io pattern issued to the file(s).
40 We may only be reading sequentially from this
41 file(s), or we may be writing randomly. Or even
42 mixing reads and writes, sequentially or randomly.
44 Block size In how large chunks are we issuing io? This may be
45 a single value, or it may describe a range of
48 IO size How much data are we going to be reading/writing.
50 IO engine How do we issue io? We could be memory mapping the
51 file, we could be using regular read/write, we
52 could be using splice, async io, syslet, or even
55 IO depth If the io engine is async, how large a queuing
56 depth do we want to maintain?
58 IO type Should we be doing buffered io, or direct/raw io?
60 Num files How many files are we spreading the workload over.
62 Num threads How many threads or processes should we spread
65 The above are the basic parameters defined for a workload, in addition
66 there's a multitude of parameters that modify other aspects of how this
72 See the README file for command line parameters, there are only a few
75 Running fio is normally the easiest part - you just give it the job file
76 (or job files) as parameters:
80 and it will start doing what the job_file tells it to do. You can give
81 more than one job file on the command line, fio will serialize the running
82 of those files. Internally that is the same as using the 'stonewall'
83 parameter described the the parameter section.
85 If the job file contains only one job, you may as well just give the
86 parameters on the command line. The command line parameters are identical
87 to the job parameters, with a few extra that control global parameters
88 (see README). For example, for the job file parameter iodepth=2, the
89 mirror command line option would be --iodepth 2 or --iodepth=2. You can
90 also use the command line for giving more than one job entry. For each
91 --name option that fio sees, it will start a new job with that name.
92 Command line entries following a --name entry will apply to that job,
93 until there are no more entries or a new --name entry is seen. This is
94 similar to the job file options, where each option applies to the current
95 job until a new [] job entry is seen.
97 fio does not need to run as root, except if the files or devices specified
98 in the job section requires that. Some other options may also be restricted,
99 such as memory locking, io scheduler switching, and decreasing the nice value.
104 As previously described, fio accepts one or more job files describing
105 what it is supposed to do. The job file format is the classic ini file,
106 where the names enclosed in [] brackets define the job name. You are free
107 to use any ascii name you want, except 'global' which has special meaning.
108 A global section sets defaults for the jobs described in that file. A job
109 may override a global section parameter, and a job file may even have
110 several global sections if so desired. A job is only affected by a global
111 section residing above it. If the first character in a line is a ';' or a
112 '#', the entire line is discarded as a comment.
114 So lets look at a really simple job file that define to threads, each
115 randomly reading from a 128MiB file.
117 ; -- start job file --
128 As you can see, the job file sections themselves are empty as all the
129 described parameters are shared. As no filename= option is given, fio
130 makes up a filename for each of the jobs as it sees fit. On the command
131 line, this job would look as follows:
133 $ fio --name=global --rw=randread --size=128m --name=job1 --name=job2
136 Lets look at an example that have a number of processes writing randomly
139 ; -- start job file --
151 Here we have no global section, as we only have one job defined anyway.
152 We want to use async io here, with a depth of 4 for each file. We also
153 increased the buffer size used to 32KiB and define numjobs to 4 to
154 fork 4 identical jobs. The result is 4 processes each randomly writing
155 to their own 64MiB file. Instead of using the above job file, you could
156 have given the parameters on the command line. For this case, you would
159 $ fio --name=random-writers --ioengine=libaio --iodepth=4 --rw=randwrite --bs=32k --direct=0 --size=64m --numjobs=4
161 fio ships with a few example job files, you can also look there for
165 5.0 Detailed list of parameters
166 -------------------------------
168 This section describes in details each parameter associated with a job.
169 Some parameters take an option of a given type, such as an integer or
170 a string. The following types are used:
172 str String. This is a sequence of alpha characters.
173 int Integer. A whole number value, may be negative.
174 siint SI integer. A whole number value, which may contain a postfix
175 describing the base of the number. Accepted postfixes are k/m/g,
176 meaning kilo, mega, and giga. So if you want to specify 4096,
177 you could either write out '4096' or just give 4k. The postfixes
178 signify base 2 values, so 1024 is 1k and 1024k is 1m and so on.
179 If the option accepts an upper and lower range, use a colon ':'
180 or minus '-' to seperate such values. See irange.
181 bool Boolean. Usually parsed as an integer, however only defined for
182 true and false (1 and 0).
183 irange Integer range with postfix. Allows value range to be given, such
184 as 1024-4096. A colon may also be used as the seperator, eg
185 1k:4k. If the option allows two sets of ranges, they can be
186 specified with a ',' or '/' delimiter: 1k-4k/8k-32k. Also see
189 With the above in mind, here follows the complete list of fio job
192 name=str ASCII name of the job. This may be used to override the
193 name printed by fio for this job. Otherwise the job
194 name is used. On the command line this parameter has the
195 special purpose of also signaling the start of a new
198 description=str Text description of the job. Doesn't do anything except
199 dump this text description when this job is run. It's
202 directory=str Prefix filenames with this directory. Used to places files
203 in a different location than "./".
205 filename=str Fio normally makes up a filename based on the job name,
206 thread number, and file number. If you want to share
207 files between threads in a job or several jobs, specify
208 a filename for each of them to override the default. If
209 the ioengine used is 'net', the filename is the host and
210 port to connect to in the format of =host/port. If the
211 ioengine is file based, you can specify a number of files
212 by seperating the names with a ':' colon. So if you wanted
213 a job to open /dev/sda and /dev/sdb as the two working files,
214 you would use filename=/dev/sda:/dev/sdb
216 opendir=str Tell fio to recursively add any file it can find in this
217 directory and down the file system tree.
220 rw=str Type of io pattern. Accepted values are:
222 read Sequential reads
223 write Sequential writes
224 randwrite Random writes
225 randread Random reads
226 rw Sequential mixed reads and writes
227 randrw Random mixed reads and writes
229 For the mixed io types, the default is to split them 50/50.
230 For certain types of io the result may still be skewed a bit,
231 since the speed may be different.
233 randrepeat=bool For random IO workloads, seed the generator in a predictable
234 way so that results are repeatable across repetitions.
236 size=siint The total size of file io for this job. This may describe
237 the size of the single file the job uses, or it may be
238 divided between the number of files in the job. If the
239 file already exists, the file size will be adjusted to this
240 size if larger than the current file size. If this parameter
241 is not given and the file exists, the file size will be used.
243 filesize=siint Individual file sizes. May be a range, in which case fio
244 will select sizes for files at random within the given range
245 and limited to 'size' in total (if that is given). If not
246 given, each created file is the same size.
249 bs=siint The block size used for the io units. Defaults to 4k. Values
250 can be given for both read and writes. If a single siint is
251 given, it will apply to both. If a second siint is specified
252 after a comma, it will apply to writes only. In other words,
253 the format is either bs=read_and_write or bs=read,write.
254 bs=4k,8k will thus use 4k blocks for reads, and 8k blocks
255 for writes. If you only wish to set the write size, you
256 can do so by passing an empty read size - bs=,8k will set
257 8k for writes and leave the read default value.
259 blocksize_range=irange
260 bsrange=irange Instead of giving a single block size, specify a range
261 and fio will mix the issued io block sizes. The issued
262 io unit will always be a multiple of the minimum value
263 given (also see bs_unaligned). Applies to both reads and
264 writes, however a second range can be given after a comma.
268 bs_unaligned If this option is given, any byte size value within bsrange
269 may be used as a block range. This typically wont work with
270 direct IO, as that normally requires sector alignment.
272 nrfiles=int Number of files to use for this job. Defaults to 1.
274 openfiles=int Number of files to keep open at the same time. Defaults to
275 the same as nrfiles, can be set smaller to limit the number
278 file_service_type=str Defines how fio decides which file from a job to
279 service next. The following types are defined:
281 random Just choose a file at random.
283 roundrobin Round robin over open files. This
286 The string can have a number appended, indicating how
287 often to switch to a new file. So if option random:4 is
288 given, fio will switch to a new random file after 4 ios
291 ioengine=str Defines how the job issues io to the file. The following
294 sync Basic read(2) or write(2) io. lseek(2) is
295 used to position the io location.
297 libaio Linux native asynchronous io.
299 posixaio glibc posix asynchronous io.
301 mmap File is memory mapped and data copied
302 to/from using memcpy(3).
304 splice splice(2) is used to transfer the data and
305 vmsplice(2) to transfer data from user
308 syslet-rw Use the syslet system calls to make
309 regular read/write async.
311 sg SCSI generic sg v3 io. May either be
312 synchronous using the SG_IO ioctl, or if
313 the target is an sg character device
314 we use read(2) and write(2) for asynchronous
317 null Doesn't transfer any data, just pretends
318 to. This is mainly used to exercise fio
319 itself and for debugging/testing purposes.
321 net Transfer over the network to given host:port.
322 'filename' must be set appropriately to
323 filename=host/port regardless of send
324 or receive, if the latter only the port
327 cpu Doesn't transfer any data, but burns CPU
328 cycles according to the cpuload= and
329 cpucycle= options. Setting cpuload=85
330 will cause that job to do nothing but burn
333 external Prefix to specify loading an external
334 IO engine object file. Append the engine
335 filename, eg ioengine=external:/tmp/foo.o
336 to load ioengine foo.o in /tmp.
338 iodepth=int This defines how many io units to keep in flight against
339 the file. The default is 1 for each file defined in this
340 job, can be overridden with a larger value for higher
343 iodepth_batch=int This defines how many pieces of IO to submit at once.
344 It defaults to the same as iodepth, but can be set lower
347 iodepth_low=int The low water mark indicating when to start filling
348 the queue again. Defaults to the same as iodepth, meaning
349 that fio will attempt to keep the queue full at all times.
350 If iodepth is set to eg 16 and iodepth_low is set to 4, then
351 after fio has filled the queue of 16 requests, it will let
352 the depth drain down to 4 before starting to fill it again.
354 direct=bool If value is true, use non-buffered io. This is usually
357 buffered=bool If value is true, use buffered io. This is the opposite
358 of the 'direct' option. Defaults to true.
360 offset=siint Start io at the given offset in the file. The data before
361 the given offset will not be touched. This effectively
362 caps the file size at real_size - offset.
364 fsync=int If writing to a file, issue a sync of the dirty data
365 for every number of blocks given. For example, if you give
366 32 as a parameter, fio will sync the file for every 32
367 writes issued. If fio is using non-buffered io, we may
368 not sync the file. The exception is the sg io engine, which
369 synchronizes the disk cache anyway.
371 overwrite=bool If writing to a file, setup the file first and do overwrites.
373 end_fsync=bool If true, fsync file contents when the job exits.
375 fsync_on_close=bool If true, fio will fsync() a dirty file on close.
376 This differs from end_fsync in that it will happen on every
377 file close, not just at the end of the job.
379 rwmixcycle=int Value in milliseconds describing how often to switch between
380 reads and writes for a mixed workload. The default is
383 rwmixread=int How large a percentage of the mix should be reads.
385 rwmixwrite=int How large a percentage of the mix should be writes. If both
386 rwmixread and rwmixwrite is given and the values do not add
387 up to 100%, the latter of the two will be used to override
390 norandommap Normally fio will cover every block of the file when doing
391 random IO. If this option is given, fio will just get a
392 new random offset without looking at past io history. This
393 means that some blocks may not be read or written, and that
394 some blocks may be read/written more than once. This option
395 is mutually exclusive with verify= for that reason.
397 nice=int Run the job with the given nice value. See man nice(2).
399 prio=int Set the io priority value of this job. Linux limits us to
400 a positive value between 0 and 7, with 0 being the highest.
403 prioclass=int Set the io priority class. See man ionice(1).
405 thinktime=int Stall the job x microseconds after an io has completed before
406 issuing the next. May be used to simulate processing being
407 done by an application. See thinktime_blocks and
411 Only valid if thinktime is set - pretend to spend CPU time
412 doing something with the data received, before falling back
413 to sleeping for the rest of the period specified by
417 Only valid if thinktime is set - control how many blocks
418 to issue, before waiting 'thinktime' usecs. If not set,
419 defaults to 1 which will make fio wait 'thinktime' usecs
422 rate=int Cap the bandwidth used by this job to this number of KiB/sec.
424 ratemin=int Tell fio to do whatever it can to maintain at least this
425 bandwidth. Failing to meet this requirement, will cause
428 rate_iops=int Cap the bandwidth to this number of IOPS. Basically the same
429 as rate, just specified independently of bandwidth. If the
430 job is given a block size range instead of a fixed value,
431 the smallest block size is used as the metric.
433 rate_iops_min=int If fio doesn't meet this rate of IO, it will cause
436 ratecycle=int Average bandwidth for 'rate' and 'ratemin' over this number
439 cpumask=int Set the CPU affinity of this job. The parameter given is a
440 bitmask of allowed CPU's the job may run on. See man
441 sched_setaffinity(2).
443 startdelay=int Start this job the specified number of seconds after fio
444 has started. Only useful if the job file contains several
445 jobs, and you want to delay starting some jobs to a certain
448 runtime=int Tell fio to terminate processing after the specified number
449 of seconds. It can be quite hard to determine for how long
450 a specified job will run, so this parameter is handy to
451 cap the total runtime to a given time.
453 invalidate=bool Invalidate the buffer/page cache parts for this file prior
454 to starting io. Defaults to true.
456 sync=bool Use sync io for buffered writes. For the majority of the
457 io engines, this means using O_SYNC.
460 mem=str Fio can use various types of memory as the io unit buffer.
461 The allowed values are:
463 malloc Use memory from malloc(3) as the buffers.
465 shm Use shared memory as the buffers. Allocated
468 shmhuge Same as shm, but use huge pages as backing.
470 mmap Use mmap to allocate buffers. May either be
471 anonymous memory, or can be file backed if
472 a filename is given after the option. The
473 format is mem=mmap:/path/to/file.
475 mmaphuge Use a memory mapped huge file as the buffer
476 backing. Append filename after mmaphuge, ala
477 mem=mmaphuge:/hugetlbfs/file
479 The area allocated is a function of the maximum allowed
480 bs size for the job, multiplied by the io depth given. Note
481 that for shmhuge and mmaphuge to work, the system must have
482 free huge pages allocated. This can normally be checked
483 and set by reading/writing /proc/sys/vm/nr_hugepages on a
484 Linux system. Fio assumes a huge page is 4MiB in size. So
485 to calculate the number of huge pages you need for a given
486 job file, add up the io depth of all jobs (normally one unless
487 iodepth= is used) and multiply by the maximum bs set. Then
488 divide that number by the huge page size. You can see the
489 size of the huge pages in /proc/meminfo. If no huge pages
490 are allocated by having a non-zero number in nr_hugepages,
491 using mmaphuge or shmhuge will fail. Also see hugepage-size.
493 mmaphuge also needs to have hugetlbfs mounted and the file
494 location should point there. So if it's mounted in /huge,
495 you would use mem=mmaphuge:/huge/somefile.
498 Defines the size of a huge page. Must at least be equal
499 to the system setting, see /proc/meminfo. Defaults to 4MiB.
500 Should probably always be a multiple of megabytes, so using
501 hugepage-size=Xm is the preferred way to set this to avoid
502 setting a non-pow-2 bad value.
504 exitall When one job finishes, terminate the rest. The default is
505 to wait for each job to finish, sometimes that is not the
508 bwavgtime=int Average the calculated bandwidth over the given time. Value
509 is specified in milliseconds.
511 create_serialize=bool If true, serialize the file creating for the jobs.
512 This may be handy to avoid interleaving of data
513 files, which may greatly depend on the filesystem
514 used and even the number of processors in the system.
516 create_fsync=bool fsync the data file after creation. This is the
519 unlink=bool Unlink the job files when done. Not the default, as repeated
520 runs of that job would then waste time recreating the fileset
523 loops=int Run the specified number of iterations of this job. Used
524 to repeat the same workload a given number of times. Defaults
527 verify=str If writing to a file, fio can verify the file contents
528 after each iteration of the job. The allowed values are:
530 md5 Use an md5 sum of the data area and store
531 it in the header of each block.
533 crc32 Use a crc32 sum of the data area and store
534 it in the header of each block.
536 This option can be used for repeated burn-in tests of a
537 system to make sure that the written data is also
540 stonewall Wait for preceeding jobs in the job file to exit, before
541 starting this one. Can be used to insert serialization
542 points in the job file. A stone wall also implies starting
543 a new reporting group.
545 new_group Start a new reporting group. If this option isn't given,
546 jobs in a file will be part of the same reporting group
547 unless seperated by a stone wall (or if it's a group
548 by itself, with the numjobs option).
550 numjobs=int Create the specified number of clones of this job. May be
551 used to setup a larger number of threads/processes doing
552 the same thing. We regard that grouping of jobs as a
555 group_reporting If 'numjobs' is set, it may be interesting to display
556 statistics for the group as a whole instead of for each
557 individual job. This is especially true of 'numjobs' is
558 large, looking at individual thread/process output quickly
559 becomes unwieldy. If 'group_reporting' is specified, fio
560 will show the final report per-group instead of per-job.
562 thread fio defaults to forking jobs, however if this option is
563 given, fio will use pthread_create(3) to create threads
566 zonesize=siint Divide a file into zones of the specified size. See zoneskip.
568 zoneskip=siint Skip the specified number of bytes when zonesize data has
569 been read. The two zone options can be used to only do
570 io on zones of a file.
572 write_iolog=str Write the issued io patterns to the specified file. See
575 read_iolog=str Open an iolog with the specified file name and replay the
576 io patterns it contains. This can be used to store a
577 workload and replay it sometime later.
579 write_bw_log If given, write a bandwidth log of the jobs in this job
580 file. Can be used to store data of the bandwidth of the
581 jobs in their lifetime. The included fio_generate_plots
582 script uses gnuplot to turn these text files into nice
585 write_lat_log Same as write_bw_log, except that this option stores io
586 completion latencies instead.
588 lockmem=siint Pin down the specified amount of memory with mlock(2). Can
589 potentially be used instead of removing memory or booting
590 with less memory to simulate a smaller amount of memory.
592 exec_prerun=str Before running this job, issue the command specified
595 exec_postrun=str After the job completes, issue the command specified
598 ioscheduler=str Attempt to switch the device hosting the file to the specified
599 io scheduler before running.
601 cpuload=int If the job is a CPU cycle eater, attempt to use the specified
602 percentage of CPU cycles.
604 cpuchunks=int If the job is a CPU cycle eater, split the load into
605 cycles of the given time. In milliseconds.
608 6.0 Interpreting the output
609 ---------------------------
611 fio spits out a lot of output. While running, fio will display the
612 status of the jobs created. An example of that would be:
614 Threads: 1: [_r] [24.8% done] [ 13509/ 8334 kb/s] [eta 00h:01m:31s]
616 The characters inside the square brackets denote the current status of
617 each thread. The possible values (in typical life cycle order) are:
621 P Thread setup, but not started.
623 I Thread initialized, waiting.
624 R Running, doing sequential reads.
625 r Running, doing random reads.
626 W Running, doing sequential writes.
627 w Running, doing random writes.
628 M Running, doing mixed sequential reads/writes.
629 m Running, doing mixed random reads/writes.
630 F Running, currently waiting for fsync()
631 V Running, doing verification of written data.
632 E Thread exited, not reaped by main thread yet.
635 The other values are fairly self explanatory - number of threads
636 currently running and doing io, rate of io since last check, and the estimated
637 completion percentage and time for the running group. It's impossible to
638 estimate runtime of the following groups (if any).
640 When fio is done (or interrupted by ctrl-c), it will show the data for
641 each thread, group of threads, and disks in that order. For each data
642 direction, the output looks like:
644 Client1 (g=0): err= 0:
645 write: io= 32MiB, bw= 666KiB/s, runt= 50320msec
646 slat (msec): min= 0, max= 136, avg= 0.03, stdev= 1.92
647 clat (msec): min= 0, max= 631, avg=48.50, stdev=86.82
648 bw (KiB/s) : min= 0, max= 1196, per=51.00%, avg=664.02, stdev=681.68
649 cpu : usr=1.49%, sys=0.25%, ctx=7969
650 IO depths : 1=0.1%, 2=0.3%, 4=0.5%, 8=99.0%, 16=0.0%, 32=0.0%, >32=0.0%
651 lat (msec): 2=1.6%, 4=0.0%, 10=3.2%, 20=12.8%, 50=38.4%, 100=24.8%,
652 lat (msec): 250=15.2%, 500=0.0%, 750=0.0%, 1000=0.0%, >=2048=0.0%
654 The client number is printed, along with the group id and error of that
655 thread. Below is the io statistics, here for writes. In the order listed,
658 io= Number of megabytes io performed
659 bw= Average bandwidth rate
660 runt= The runtime of that thread
661 slat= Submission latency (avg being the average, stdev being the
662 standard deviation). This is the time it took to submit
663 the io. For sync io, the slat is really the completion
664 latency, since queue/complete is one operation there.
665 clat= Completion latency. Same names as slat, this denotes the
666 time from submission to completion of the io pieces. For
667 sync io, clat will usually be equal (or very close) to 0,
668 as the time from submit to complete is basically just
669 CPU time (io has already been done, see slat explanation).
670 bw= Bandwidth. Same names as the xlat stats, but also includes
671 an approximate percentage of total aggregate bandwidth
672 this thread received in this group. This last value is
673 only really useful if the threads in this group are on the
674 same disk, since they are then competing for disk access.
675 cpu= CPU usage. User and system time, along with the number
676 of context switches this thread went through.
677 IO depths= The distribution of io depths over the job life time. The
678 numbers are divided into powers of 2, so for example the
679 16= entries includes depths up to that value but higher
680 than the previous entry. In other words, it covers the
682 IO latencies= The distribution of IO completion latencies. This is the
683 time from when IO leaves fio and when it gets completed.
684 The numbers follow the same pattern as the IO depths,
685 meaning that 2=1.6% means that 1.6% of the IO completed
686 within 2 msecs, 20=12.8% means that 12.8% of the IO
687 took more than 10 msecs, but less than (or equal to) 20 msecs.
689 After each client has been listed, the group statistics are printed. They
692 Run status group 0 (all jobs):
693 READ: io=64MiB, aggrb=22178, minb=11355, maxb=11814, mint=2840msec, maxt=2955msec
694 WRITE: io=64MiB, aggrb=1302, minb=666, maxb=669, mint=50093msec, maxt=50320msec
696 For each data direction, it prints:
698 io= Number of megabytes io performed.
699 aggrb= Aggregate bandwidth of threads in this group.
700 minb= The minimum average bandwidth a thread saw.
701 maxb= The maximum average bandwidth a thread saw.
702 mint= The smallest runtime of the threads in that group.
703 maxt= The longest runtime of the threads in that group.
705 And finally, the disk statistics are printed. They will look like this:
707 Disk stats (read/write):
708 sda: ios=16398/16511, merge=30/162, ticks=6853/819634, in_queue=826487, util=100.00%
710 Each value is printed for both reads and writes, with reads first. The
713 ios= Number of ios performed by all groups.
714 merge= Number of merges io the io scheduler.
715 ticks= Number of ticks we kept the disk busy.
716 io_queue= Total time spent in the disk queue.
717 util= The disk utilization. A value of 100% means we kept the disk
718 busy constantly, 50% would be a disk idling half of the time.
724 For scripted usage where you typically want to generate tables or graphs
725 of the results, fio can output the results in a semicolon separated format.
726 The format is one long line of values, such as:
728 client1;0;0;1906777;1090804;1790;0;0;0.000000;0.000000;0;0;0.000000;0.000000;929380;1152890;25.510151%;1078276.333333;128948.113404;0;0;0;0;0;0.000000;0.000000;0;0;0.000000;0.000000;0;0;0.000000%;0.000000;0.000000;100.000000%;0.000000%;324;100.0%;0.0%;0.0%;0.0%;0.0%;0.0%;0.0%;100.0%;0.0%;0.0%;0.0%;0.0%;0.0%
729 ;0.0%;0.0%;0.0%;0.0%;0.0%
731 Split up, the format is as follows:
733 jobname, groupid, error
735 KiB IO, bandwidth (KiB/sec), runtime (msec)
736 Submission latency: min, max, mean, deviation
737 Completion latency: min, max, mean, deviation
738 Bw: min, max, aggregate percentage of total, mean, deviation
740 KiB IO, bandwidth (KiB/sec), runtime (msec)
741 Submission latency: min, max, mean, deviation
742 Completion latency: min, max, mean, deviation
743 Bw: min, max, aggregate percentage of total, mean, deviation
744 CPU usage: user, system, context switches
745 IO depths: <=1, 2, 4, 8, 16, 32, >=64
746 IO latencies: <=2, 4, 10, 20, 50, 100, 250, 500, 750, 1000, >=2000