Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 22 Oct 2021 15:31:58 +0000
From:      bugzilla-noreply@freebsd.org
To:        bugs@FreeBSD.org
Subject:   [Bug 258204] Extremely poor performance copying directory with lot of small files on nvme
Message-ID:  <bug-258204-227-HD25K7COPG@https.bugs.freebsd.org/bugzilla/>
In-Reply-To: <bug-258204-227@https.bugs.freebsd.org/bugzilla/>
References:  <bug-258204-227@https.bugs.freebsd.org/bugzilla/>

next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D258204

--- Comment #3 from Ali Abdallah <ali.abdallah@suse.com> ---
Some nvme perf test measurements=20

# nvmecontrol perftest -n 1 -o read -s 4096 -t 30 nvme0ns1
Threads:  1 Size:   4096  READ Time:  30 IO/s:   34515 MB/s:  134

# nvmecontrol perftest -n 2 -o read -s 4096 -t 30 nvme0ns1
Threads:  2 Size:   4096  READ Time:  30 IO/s:   69473 MB/s:  271

diskinfo -tv /dev/nvd0p4.eli=20
/dev/nda0p4.eli
        4096            # sectorsize
        478888849408    # mediasize in bytes (446G)
        116916223       # mediasize in sectors
        0               # stripesize
        0               # stripeoffset
        7277            # Cylinders according to firmware.
        255             # Heads according to firmware.
        63              # Sectors according to firmware.
        UMIS LENSE40512GMSP34MESTB3A    # Disk descr.
        SS0L25153X3RC03Y11N2    # Disk ident.
        nvme0           # Attachment
        Yes             # TRIM/UNMAP support
        0               # Rotation rate in RPM

Seek times:
        Full stroke:      250 iter in   0.059140 sec =3D    0.237 msec
        Half stroke:      250 iter in   0.038418 sec =3D    0.154 msec
        Quarter stroke:   500 iter in   0.075839 sec =3D    0.152 msec
        Short forward:    400 iter in   0.059095 sec =3D    0.148 msec
        Short backward:   400 iter in   0.076297 sec =3D    0.191 msec
        Seq outer:       2048 iter in   0.220371 sec =3D    0.108 msec
        Seq inner:       2048 iter in   0.312280 sec =3D    0.152 msec

Transfer rates:
        outside:       102400 kbytes in   1.339011 sec =3D    76474 kbytes/=
sec
        middle:        102400 kbytes in   1.125152 sec =3D    91010 kbytes/=
sec
        inside:        102400 kbytes in   1.394730 sec =3D    73419 kbytes/=
sec

The values look really bad for a NVME device.

--=20
You are receiving this mail because:
You are the assignee for the bug.=



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-258204-227-HD25K7COPG>