Re: HyperThreading
--0023547c8ba544b37c049c96520e
Content-Type: text/plain; charset=UTF-8
On Fri, Feb 18, 2011 at 3:48 PM, Saman Barghi <saman.b@gmail.com> wrote:
>
> Could you try -master? There has been fairly substantial work since the 2.8
>> release; all of the tokens in 2.8 are still taking the MP lock as a
>> precaution. That is no longer the case in master. The process token has been
>> broken up fairly dramatically in master as well, so I expect to see
>> different results.
>>
>
> Sorry, I am not following, by master you mean the master branch !?
>
Yep; we produce CD snapshots each night, they're from the latest git code as
of that day.
>
>> Can you give us a bit more information about the h/w and configuration? In
>> Kris Kennaway's tests, he was seeing FreeBSD 7 sustain 8 times the number of
>> transactions you're reporting for 8.1, and that was for hardware that is
>> three or four years old at this point.
>>
>
> I see, good point. The machine I am testing on is a 4 core xeon with 2GB
> of RAM, and I it is 2 to 3 years old as well (The machine Kris were using
> was a "2.0Ghz 8-core amd64 system with 16GB of RAM", so as it has twice the
> number of cores and 8 times more memory than what I have here, I think the
> results are fair for number of transactions). I might be able to re run the
> tests on another machine if I get lucky. MySQL configuration is exactly the
> same as what Kris used (http://people.freebsd.org/~kris/scaling/my.cnf), I
> am using MySQL 5.0.91. Here is the sysbench command :
>
http://people.freebsd.org/~kris/scaling/4cpu-mysql.png
was on a 4 x p3
700mhz; I believe (but we should ask Kris) that the opterons were not used
for the cross-OS comparisons; they were used for the fbsd 5.5 vs 6.2 vs 7.0
tests. (see the titles on that graph vs the cross-OS comparison). Happen to
know what sort of Xeons exactly?
> sysbench --test=oltp --num-threads=${i} --mysql-user=root --max-time=120
> --max-requests=0 --oltp-read-only=on run
>
> Also, I am not sure if Kris is talking about queries/sec or
> transactions/sec, since he used both of them in two different graphs he
> presented. What I presented is the number of transactions as reported by
> sysbench, however the total number of queries are much more than this, take
> a look at this: http://www.cs.uwaterloo.ca/~sbarghi/bench/queries1.jpg
>
Ah, okay; I only used Transactions/sec.
> Also, last I tried sysbench/OLTP with MySQL on dfly, I saw MySQL making
>> many 256K allocations, which fall back to mmap() in our system. mmap()
>> acquires the VM token, so I expected to see a much more dramatic falloff
>> than you saw.
>>
>
> I will double check if everything is working properly and there is no
> mistake in configuration of both machines. Do you remember which version of
> MySQL you were using ?and What is the best way of profiling in DragonFly ?!
>
I used MySQL 5.1.37 I believe. Watch the output of systat -pv; you'll be
able to see contended tokens.
Thanks!
-- vs
--0023547c8ba544b37c049c96520e
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
<div class=3D"gmail_quote">On Fri, Feb 18, 2011 at 3:48 PM, Saman Barghi <s=
pan dir=3D"ltr"><<a href=3D"mailto:saman.b@gmail.com">saman.b@gmail.com<=
/a>></span> wrote:<br><blockquote class=3D"gmail_quote" style=3D"border-=
left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left=
: 1ex;">
<div dir=3D"ltr"><br><div class=3D"gmail_quote"><div class=3D"im"><blockquo=
te class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, 204, 204)=
; margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"><div class=3D"gmail_quote"=
><div>Could you try -master? There has been fairly substantial work since t=
he 2.8 release; all of the tokens in 2.8 are still taking the MP lock as a =
precaution. That is no longer the case in master. The process token has bee=
n broken up fairly dramatically in master as well, so I expect to see diffe=
rent results.<br>
</div></div></blockquote><div>=C2=A0</div></div><div>Sorry, I am not follow=
ing, by master you mean the master branch !? <br></div></div></div></blockq=
uote><div><br>Yep; we produce CD snapshots each night, they're from the=
latest git code as of that day.<br>
=C2=A0</div><blockquote class=3D"gmail_quote" style=3D"border-left: 1px sol=
id rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"><div =
dir=3D"ltr"><div class=3D"gmail_quote"><div></div><div class=3D"im"><blockq=
uote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, 204, 20=
4); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<div class=3D"gmail_quote"><div>
<br>Can you give us a bit more information about the h/w and configuration?=
In Kris Kennaway's tests, he was seeing FreeBSD 7 sustain 8 times the =
number of transactions you're reporting for 8.1, and that was for hardw=
are that is three or four years old at this point. </div>
</div></blockquote></div><div><br>I see, good point. The machine I am testi=
ng on=C2=A0 is a=C2=A0 4 core xeon with 2GB of RAM, and I it
is=C2=A0 2 to 3 years old as well (The machine Kris were using was a "=
;2.0Ghz
8-core amd64 system with 16GB of
RAM", so as it has twice the number of cores and 8 times more memory=
=20
than what I have here, I think the results are fair for number of=20
transactions). I might be able to re run the tests on another=20
machine if I get lucky. MySQL=20
configuration is exactly the same as what Kris used=20
(<a href=3D"http://people.freebsd.org/%7Ekris/scaling/my.cnf" target=3D"_bl=
ank">http://people.freebsd.org/~kris/scaling/my.cnf</a>), I am using MySQL =
5.0.91. Here is the sysbench command : <br></div></div></div></blockquote>
<br><a href=3D"http=" rel="nofollow">http://people.freebsd.org/~kris/scaling/4cpu-mysql.png">http=
://people.freebsd.org/~kris/scaling/4cpu-mysql.png</a> was on a 4 x p3 700m=
hz; I believe (but we should ask Kris) that the opterons were not used for =
the cross-OS comparisons; they were used for the fbsd 5.5 vs 6.2 vs 7.0 tes=
ts. (see the titles on that graph vs the cross-OS comparison). Happen to kn=
ow what sort of Xeons exactly?<br>
<div>=C2=A0</div><blockquote class=3D"gmail_quote" style=3D"border-left: 1p=
x solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">=
<div dir=3D"ltr"><div class=3D"gmail_quote"><div>
sysbench --test=3Doltp --num-threads=3D${i} --mysql-user=3Droot --max-time=
=3D120
--max-requests=3D0 --oltp-read-only=3Don run<br><br>Also, I am not sure if =
Kris is talking about queries/sec or=20
transactions/sec, since he used both of them in two different graphs he=20
presented. What I presented is the number of transactions as reported by
sysbench, however the total number of queries are much more than this,=20
take a look at this:=20
<a href=3D"http://www.cs.uwaterloo.ca/%7Esbarghi/bench/queries1.jpg" target=
=3D"_blank">http://www.cs.uwaterloo.ca/~sbarghi/bench/queries1.jpg</a> <br>=
</div></div></div></blockquote><div><br>Ah, okay; I only used Transactions/=
sec.<br>
=C2=A0</div><blockquote class=3D"gmail_quote" style=3D"border-left: 1px sol=
id rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"><div =
dir=3D"ltr"><div class=3D"gmail_quote"><div></div><div class=3D"im"><blockq=
uote class=3D"gmail_quote" style=3D"border-left: 1px solid rgb(204, 204, 20=
4); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<div class=3D"gmail_quote"><div>Also, last I tried sysbench/OLTP with MySQL=
on dfly, I saw MySQL making many 256K allocations, which fall back to mmap=
() in our system. mmap() acquires the VM token, so I expected to see a much=
more dramatic falloff than you saw.<br>
</div></div></blockquote></div><div>=C2=A0<br>I will double check if everyt=
hing is working properly and there is no mistake in configuration of both m=
achines. Do you remember which version of MySQL you were using ?and What is=
the best way of profiling in DragonFly ?! <br>
</div></div></div></blockquote><div><br>I used MySQL 5.1.37 I believe. Watc=
h the output of systat -pv; you'll be able to see contended tokens.<br>=
<br>Thanks!<br>-- vs<br></div></div>
--0023547c8ba544b37c049c96520e--
討論串 (同標題文章)
完整討論串 (本文為第 13 之 16 篇)：