Re: [PATCH 0/6] RFC v2: mm: gup/dma tracking

[Tom Talpey]

On 2/5/2019 3:22 AM, John Hubbard wrote:
> On 2/4/19 5:41 PM, Tom Talpey wrote:
> > On 2/4/2019 12:21 AM, john.hubbard@xxxxxxxxx wrote:
> > > From: John Hubbard <jhubbard@xxxxxxxxxx>
> > >
> > >
> > > Performance: here is an fio run on an NVMe drive, using this for the fio
> > > configuration file:
> > >
> > > ÂÂÂÂ [reader]
> > > ÂÂÂÂ direct=1
> > > ÂÂÂÂ ioengine=libaio
> > > ÂÂÂÂ blocksize=4096
> > > ÂÂÂÂ size=1g
> > > ÂÂÂÂ numjobs=1
> > > ÂÂÂÂ rw=read
> > > ÂÂÂÂ iodepth=64
> > >
> > > reader: (g=0): rw=read, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T) 
> > > 4096B-4096B, ioengine=libaio, iodepth=64
> > > fio-3.3
> > > Starting 1 process
> > > Jobs: 1 (f=1)
> > > reader: (groupid=0, jobs=1): err= 0: pid=7011: Sun Feb 3 20:36:51 2019
> > > ÂÂÂ read: IOPS=190k, BW=741MiB/s (778MB/s)(1024MiB/1381msec)
> > > ÂÂÂÂ slat (nsec): min=2716, max=57255, avg=4048.14, stdev=1084.10
> > > ÂÂÂÂ clat (usec): min=20, max=12485, avg=332.63, stdev=191.77
> > > ÂÂÂÂÂ lat (usec): min=22, max=12498, avg=336.72, stdev=192.07
> > > ÂÂÂÂ clat percentiles (usec):
> > >  | 1.00th=[ 322], 5.00th=[ 322], 10.00th=[ 322], 20.00th=[ 
> > > 326],
> > > ÂÂÂÂÂ | 30.00th=[Â 326], 40.00th=[Â 326], 50.00th=[Â 326], 60.00th=[ 
> > > 326],
> > > ÂÂÂÂÂ | 70.00th=[Â 326], 80.00th=[Â 330], 90.00th=[Â 330], 95.00th=[ 
> > > 330],
> > > ÂÂÂÂÂ | 99.00th=[Â 478], 99.50th=[Â 717], 99.90th=[ 1074], 99.95th=[ 
> > > 1090],
> > > ÂÂÂÂÂ | 99.99th=[12256]
> >
> > These latencies are concerning. The best results we saw at the end of
> > November (previous approach) were MUCH flatter. These really start
> > spiking at three 9's, and are sky-high at four 9's. The "stdev" values
> > for clat and lat are about 10 times the previous. There's some kind
> > of serious queuing contention here, that wasn't there in November.
>
> Hi Tom,
>
> I think this latency problem is also there in the baseline kernel, but...
>
> > > ÂÂÂ bw (Â KiB/s): min=730152, max=776512, per=99.22%, avg=753332.00, 
> > > stdev=32781.47, samples=2
> > > ÂÂÂ iopsÂÂÂÂÂÂÂ : min=182538, max=194128, avg=188333.00, 
> > > stdev=8195.37, samples=2
> > > ÂÂ lat (usec)ÂÂ : 50=0.01%, 100=0.01%, 250=0.07%, 500=99.26%, 750=0.38%
> > > ÂÂ lat (usec)ÂÂ : 1000=0.02%
> > > ÂÂ lat (msec)ÂÂ : 2=0.24%, 20=0.02%
> > > ÂÂ cpuÂÂÂÂÂÂÂÂÂ : usr=15.07%, sys=84.13%, ctx=10, majf=0, minf=74
> >
> > System CPU 84% is roughly double the November results of 45%. Ouch.
>
> That's my fault. First of all, I had a few extra, supposedly minor debug
> settings in the .config, which I'm removing now--I'm doing a proper run
> with the original .config file from November, below. Second, I'm not
> sure I controlled the run carefully enough.
>
> > Did you re-run the baseline on the new unpatched base kernel and can
> > we see the before/after?
>
> Doing that now, I see:
>
> -- No significant perf difference between before and after, but
> -- Still high clat in the 99.99th
>
> =======================================================================
> Before: using commit 8834f5600cf3 ("Linux 5.0-rc5")
> ===================================================
> reader: (g=0): rw=read, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T) 
> 4096B-4096B, ioengine=libaio, iodepth=64
> fio-3.3
> Starting 1 process
> Jobs: 1 (f=1)
> reader: (groupid=0, jobs=1): err= 0: pid=1829: Tue Feb 5 00:08:08 2019
>  ÂÂ read: IOPS=193k, BW=753MiB/s (790MB/s)(1024MiB/1359msec)
>  ÂÂÂ slat (nsec): min=1269, max=40309, avg=1493.66, stdev=534.83
>  ÂÂÂ clat (usec): min=127, max=12249, avg=329.83, stdev=184.92
>  ÂÂÂÂ lat (usec): min=129, max=12256, avg=331.35, stdev=185.06
>  ÂÂÂ clat percentiles (usec):
>   | 1.00th=[ 326], 5.00th=[ 326], 10.00th=[ 326], 20.00th=[ 326],
>  ÂÂÂÂ | 30.00th=[Â 326], 40.00th=[Â 326], 50.00th=[Â 326], 60.00th=[Â 326],
>  ÂÂÂÂ | 70.00th=[Â 326], 80.00th=[Â 326], 90.00th=[Â 326], 95.00th=[Â 326],
>  ÂÂÂÂ | 99.00th=[Â 347], 99.50th=[Â 519], 99.90th=[Â 529], 99.95th=[Â 537],
>  ÂÂÂÂ | 99.99th=[12125]
>  ÂÂ bw (Â KiB/s): min=755032, max=781472, per=99.57%, avg=768252.00, 
> stdev=18695.90, samples=2
>  ÂÂ iopsÂÂÂÂÂÂÂ : min=188758, max=195368, avg=192063.00, stdev=4673.98, 
> samples=2
>  Â lat (usec)ÂÂ : 250=0.08%, 500=99.18%, 750=0.72%
>  Â lat (msec)ÂÂ : 20=0.02%
>  Â cpuÂÂÂÂÂÂÂÂÂ : usr=12.30%, sys=46.83%, ctx=253554, majf=0, minf=74
>  Â IO depthsÂÂÂ : 1=0.1%, 2=0.1%, 4=0.1%, 8=0.1%, 16=0.1%, 32=0.1%, 
>  >=64=100.0%
>  ÂÂÂÂ submitÂÂÂ : 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=0.1%, 
>  >=64=0.0%
>  ÂÂÂÂ issued rwts: total=262144,0,0,0 short=0,0,0,0 dropped=0,0,0,0
>  ÂÂÂÂ latencyÂÂ : target=0, window=0, percentile=100.00%, depth=64
>
> Run status group 0 (all jobs):
>  ÂÂ READ: bw=753MiB/s (790MB/s), 753MiB/s-753MiB/s (790MB/s-790MB/s), 
> io=1024MiB (1074MB), run=1359-1359msec
>
> Disk stats (read/write):
>  Â nvme0n1: ios=221246/0, merge=0/0, ticks=71556/0, in_queue=704, 
> util=91.35%
>
> =======================================================================
> After:
> =======================================================================
> reader: (g=0): rw=read, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T) 
> 4096B-4096B, ioengine=libaio, iodepth=64
> fio-3.3
> Starting 1 process
> Jobs: 1 (f=1)
> reader: (groupid=0, jobs=1): err= 0: pid=1803: Mon Feb 4 23:58:07 2019
>  ÂÂ read: IOPS=193k, BW=753MiB/s (790MB/s)(1024MiB/1359msec)
>  ÂÂÂ slat (nsec): min=1276, max=41900, avg=1505.36, stdev=565.26
>  ÂÂÂ clat (usec): min=177, max=12186, avg=329.88, stdev=184.03
>  ÂÂÂÂ lat (usec): min=178, max=12192, avg=331.42, stdev=184.16
>  ÂÂÂ clat percentiles (usec):
>   | 1.00th=[ 326], 5.00th=[ 326], 10.00th=[ 326], 20.00th=[ 326],
>  ÂÂÂÂ | 30.00th=[Â 326], 40.00th=[Â 326], 50.00th=[Â 326], 60.00th=[Â 326],
>  ÂÂÂÂ | 70.00th=[Â 326], 80.00th=[Â 326], 90.00th=[Â 326], 95.00th=[Â 326],
>  ÂÂÂÂ | 99.00th=[Â 359], 99.50th=[Â 498], 99.90th=[Â 537], 99.95th=[Â 627],
>  ÂÂÂÂ | 99.99th=[12125]
>  ÂÂ bw (Â KiB/s): min=754656, max=781504, per=99.55%, avg=768080.00, 
> stdev=18984.40, samples=2
>  ÂÂ iopsÂÂÂÂÂÂÂ : min=188664, max=195378, avg=192021.00, stdev=4747.51, 
> samples=2
>  Â lat (usec)ÂÂ : 250=0.12%, 500=99.40%, 750=0.46%
>  Â lat (msec)ÂÂ : 20=0.02%
>  Â cpuÂÂÂÂÂÂÂÂÂ : usr=12.44%, sys=47.05%, ctx=252127, majf=0, minf=73
>  Â IO depthsÂÂÂ : 1=0.1%, 2=0.1%, 4=0.1%, 8=0.1%, 16=0.1%, 32=0.1%, 
>  >=64=100.0%
>  ÂÂÂÂ submitÂÂÂ : 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=0.1%, 
>  >=64=0.0%
>  ÂÂÂÂ issued rwts: total=262144,0,0,0 short=0,0,0,0 dropped=0,0,0,0
>  ÂÂÂÂ latencyÂÂ : target=0, window=0, percentile=100.00%, depth=64
>
> Run status group 0 (all jobs):
>  ÂÂ READ: bw=753MiB/s (790MB/s), 753MiB/s-753MiB/s (790MB/s-790MB/s), 
> io=1024MiB (1074MB), run=1359-1359msec
>
> Disk stats (read/write):
>  Â nvme0n1: ios=221203/0, merge=0/0, ticks=71291/0, in_queue=704, 
> util=91.19%
>
> How's this look to you?

Ok, I'm satisfied the four-9's latency spike is in not your code. :-)
Results look good relative to baseline. Thanks for doublechecking!

Tom.