Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 22 Feb 2019 07:51:44 -0800
From:      Enji Cooper <yaneurabeya@gmail.com>
To:        Rajesh Kumar <rajfbsd@gmail.com>
Cc:        FreeBSD Hackers <freebsd-hackers@freebsd.org>
Subject:   Re: Any ideal way to run FIO benchmarking for NVMEe devices in FreeBSD
Message-ID:  <DEF69F8F-3CA0-448C-BA02-52A22CE8ECAB@gmail.com>
In-Reply-To: <CAAO%2BANPUF8u54Xvu4uN=nRAEXaLNtBX%2BJkYzRbb5jrZMRTZ6gA@mail.gmail.com>
References:  <CAAO%2BANM34aY4g%2BFjPdt8F2sNo5e6N2dZdTDKavEJwvRbNJz=Gw@mail.gmail.com> <0E136DED-C1AD-481C-B243-C943D4F8D9C5@gmail.com> <CAAO%2BANPUF8u54Xvu4uN=nRAEXaLNtBX%2BJkYzRbb5jrZMRTZ6gA@mail.gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help

> On Feb 22, 2019, at 01:29, Rajesh Kumar <rajfbsd@gmail.com> wrote:
>=20
> Hi Enji Cooper,
>=20
> I am using Samsung 960 PRO
>=20
> https://www.samsung.com/semiconductor/minisite/ssd/product/consumer/960pro=
/

Hi Rajesh,
    I asked about the datasheet, because there might be some hints in terms o=
f the number of parallel jobs you might want to apply as well as the I/O que=
ue depth, which will affect the performance of the drive. Otherwise you=E2=80=
=99ll be throwing values against a wall, seeing what will stick, which is so=
rt of ok if you bound your testing and adjust based on performance, but if y=
our initial hunch is off, it can mislead you.
    Similarly, check to see if there are any tunables or sysctls that will b=
ound the device in terms of the queue depth and I/O requests.
    As others have noted, test the device directly if you want to know its r=
aw performance. Only test with a filesystem if your intent is to see how the=
 device will perform with a filesystem on it (and disable filesystem sync if=
 you want to test max throughput with the overhead of a filesystem). Testing=
 with a filesystem can reveal some potentially interesting characteristics, i=
n terms of limitations in VFS / the filesystem implementation, which might b=
e helpful if you=E2=80=99re trying to determine why there=E2=80=99s a differ=
ence between raw device speed and the speed with a filesystem on it. Testing=
 with the same file in different directories is ok, as long as you blow the d=
rive cache=E2=80=94which will have a noticeable performance impact on conven=
tional (PMR, SMR, etc) drives, as you want to test the worst case speed of t=
he device, instead of the cache. It should matter a bit less with faster dev=
ices like SSDs/NVMe drives. Testing with files in the same lateral filesyste=
m hierarchy (same directory), might reveal issues with filesystem locking (d=
irectory inode performance), but that shouldn=E2=80=99t be the primary goal o=
f your testing. It=E2=80=99s just something to keep in mind.
    Happy testing!
HTH,
-Enji=



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?DEF69F8F-3CA0-448C-BA02-52A22CE8ECAB>