Re: was: configured swap now: relative rpi2+ ufs2 nfs random rw speeds

On Wed, Nov 12, 2025 at 12:39:46PM +0000, void wrote:
> On Wed, Nov 12, 2025 at 01:10:31AM +0000, void wrote:
> 
> > The rpi2b+ is busy rn building fio then it needs to
> > re-build+installkernel. I'll run local and nfs testing
> > afterwards then report back
> 
> fio won't build with libnfs option and it appears to need that
> to test performance of nfs mounts. I've raised a PR
> https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=290971
> -- 

Here's the result of an fio run on a Pi2 running
FreeBSD generic 16.0-CURRENT FreeBSD 16.0-CURRENT #2 main-n281742-3f5f52216f7e: Tue Nov 11 11:40:31 PST 2025     root@generic:/usr/obj/usr/src/arm.armv7/sys/GENERIC arm

 
root@generic:/usr/ports/benchmarks/fio # fio --name=random_rw_test --filename=./testfile1 --rw=randrw --bs=128k \
> --ioengine=posixaio --iodepth=256 --numjobs=4 --runtime=120 --time_based \
reporting --> direct=1 --size=1G--group_reporting --direct=1 --size=1G
random_rw_test: (g=0): rw=randrw, bs=(R) 128KiB-128KiB, (W) 128KiB-128KiB, (T) 128KiB-128KiB, ioengine=posixaio, iodepth=256
...
fio-3.40
Starting 4 processes
random_rw_test: Laying out IO file (1 file / 1024MiB)
Jobs: 3 (f=3): [f(1),_(1),f(1),m(1)][54.3%][r=15.9MiB/s,w=13.6MiB/s][r=126,w=108 IOPS][eta 01m:Jobs: 1 (f=1): [_(3),m(1)][33.3%][eta 04m:30s]                                                     
random_rw_test: (groupid=0, jobs=4): err= 0: pid=32259: Wed Nov 12 08:35:36 2025
  read: IOPS=43, BW=5630KiB/s (5765kB/s)(736MiB/133780msec)
    slat (usec): min=15, max=23806, avg=30.71, stdev=311.25
    clat (msec): min=1503, max=31007, avg=11260.63, stdev=3607.37
     lat (msec): min=1503, max=31007, avg=11260.66, stdev=3607.36
    clat percentiles (msec):
     |  1.00th=[ 1519],  5.00th=[ 3943], 10.00th=[ 7617], 20.00th=[ 9731],
     | 30.00th=[10268], 40.00th=[10805], 50.00th=[11342], 60.00th=[11879],
     | 70.00th=[12416], 80.00th=[13087], 90.00th=[14160], 95.00th=[16442],
     | 99.00th=[17113], 99.50th=[17113], 99.90th=[17113], 99.95th=[17113],
     | 99.99th=[17113]
   bw (  KiB/s): min= 2043, max=117956, per=100.00%, avg=37211.27, stdev=8970.88, samples=161
   iops        : min=   13, max=  921, avg=289.16, stdev=70.15, samples=161
  write: IOPS=45, BW=5880KiB/s (6021kB/s)(768MiB/133780msec); 0 zone resets
    slat (usec): min=43, max=2553, avg=117.10, stdev=75.50
    clat (msec): min=1501, max=25701, avg=10672.64, stdev=2908.87
     lat (msec): min=1501, max=25701, avg=10672.76, stdev=2908.87
    clat percentiles (msec):
     |  1.00th=[ 1502],  5.00th=[ 3943], 10.00th=[ 6879], 20.00th=[ 9463],
     | 30.00th=[10134], 40.00th=[10402], 50.00th=[11208], 60.00th=[11610],
     | 70.00th=[12013], 80.00th=[12818], 90.00th=[13489], 95.00th=[14160],
     | 99.00th=[17113], 99.50th=[17113], 99.90th=[17113], 99.95th=[17113],
     | 99.99th=[17113]
   bw (  KiB/s): min= 1021, max=140151, per=100.00%, avg=52665.09, stdev=12639.90, samples=120
   iops        : min=    5, max= 1094, avg=409.93, stdev=98.83, samples=120
  lat (msec)   : 2000=2.04%, >=2000=97.96%
  cpu          : usr=0.17%, sys=0.92%, ctx=11512, majf=3, minf=20
  IO depths    : 1=0.1%, 2=0.1%, 4=0.1%, 8=0.3%, 16=0.6%, 32=4.0%, >=64=94.9%
     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=98.1%, 8=0.6%, 16=0.4%, 32=0.3%, 64=0.1%, >=64=0.4%
     issued rwts: total=5884,6145,0,0 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=256

Run status group 0 (all jobs):
   READ: bw=5630KiB/s (5765kB/s), 5630KiB/s-5630KiB/s (5765kB/s-5765kB/s), io=736MiB (771MB), run=133780-133780msec
  WRITE: bw=5880KiB/s (6021kB/s), 5880KiB/s-5880KiB/s (6021kB/s-6021kB/s), io=768MiB (805MB), run=133780-133780msec
root@generic:/usr/ports/benchmarks/fio # 

The filesystem us ufs, on a usb mechanical hard disk.

Thanks for bringing fio to light!

bob prohaska