Re: Implementation PrevaylerEight
"Flavio W. Brasil" <[email protected]> Tue, 21 Oct 2014 23:01:05 +0200
| Newsgroups | gmane.comp.java.prevayler |
|---|---|
| Message-ID | <[email protected]> |
--===============3913037630437737279==
Content-Type: multipart/alternative; boundary="5446c991_dcdf8f6_104"
--5446c991_dcdf8f6_104
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
Hi Christian, =20
Thanks for sharing Prevayler8, it looks promising=21 I tried to implement=
=20
something similar in the past, but didn=E2=80=99t have time to finish.
I think http://databen.ch could be a good tool to test the scalability =20
and correctness of the implementation. I can help with it if you want.
Cheers, =20
-- =20
=46lavio W. Brasil
On Tuesday, October 21, 2014 at 10:34 PM, Christian Stein wrote:
> =20
> =20
> On Tue, Oct 21, 2014 at 9:25 PM, Marcel Oerlemans <marceloerlemans=40gm=
ail.com (mailto:marceloerlemans=40gmail.com)> wrote:
> > Hi Christian,
> > =20
> > Those numbers are impressive=21
> > I see you tested default Prevayler up to only 8 threads. You might wa=
nt to scale that up a bit=3F =20
> > =20
> =20
> =20
> Did this already and it peaked at about 900 threads here with 18k opera=
tions per second.
> =20
> > After I read your mail, I ran a prevayler scalability test, from 1 to=
1500 threads, 20 seconds per round/
> > In my test setup Prevayler handled 60 transactions/second =40 8 threa=
ds, but at 1489 threads, Prevayler could handle 5945 transactions / secon=
d. That's =7E100 times as many. Results in graph form can be found here: =
http://imgur.com/sRIaCVD
> > =20
> =20
> =20
> What machine are you running the test on=3F
> =20
> > I wanted to do a similar test with Prevayler8, but I see now I need t=
wo classes which I cant seem to find: ByteBufferOutputStream and ByteBuff=
erInputStream. Can you help me with that=3F
> > =20
> =20
> See the first mail - and mind the purpose and limitation of that implem=
entation: shines at single-threaded environments (no sync) and sacrifices=
some Durability points to speed. I'm playing with memory.force() at the =
moment and will integrate that soon.
> =20
> What I'm going to test next, is the default implementation configured f=
or speed: no deep copy, no forced disk sync.
> =20
> And draw some charts, too. ;-)
> =20
> Cheers,
> Christian
> =20
> =20
> =20
> =20
> -----------------------------------------------------------------------=
-------
> Comprehensive Server Monitoring with Site24x7.
> Monitor 10 servers for =249/Month.
> Get alerted through email, SMS, voice calls or mobile push notification=
s.
> Take corrective actions from your mobile device.
> http://p.sf.net/sfu/Zoho
> =20
> =5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=
=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F
> To unsubscribe go to the end of this page: http://lists.sourceforge.net=
/lists/listinfo/prevayler-discussion
> =5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=
=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F
> =22Databases in Memoriam=22 -- http://www.prevayler.org
> =20
> =20
--5446c991_dcdf8f6_104
Content-Type: text/html; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
<div>
Hi Christian,
</div><div><br></div><div>Thanks for sharing Prevayler8, =
it looks promising=21 I tried to implement </div><div>something simi=
lar in the past, but didn=E2=80=99t have time to finish.</div><div><br></=
div><div>I think http://databen.ch could be a good tool to test the scala=
bility </div><div>and correctness of the implementation. I can help =
with it if you want.</div><div><br></div><div>Cheers,</div>
<div><div><br></div><div>-- </div><div>=46lavio W. B=
rasil<br></div><div><br></div></div>
=20
<p style=3D=22color: =23A0A0A8;=22>On Tuesday, October 21=
, 2014 at 10:34 PM, Christian Stein wrote:</p>
<blockquote type=3D=22cite=22 style=3D=22border-left-styl=
e:solid;border-width:1px;margin-left:0px;padding-left:10px;=22>
<span><div><div><div dir=3D=22ltr=22><br><div><br><di=
v>On Tue, Oct 21, 2014 at 9:25 PM, Marcel Oerlemans <span dir=3D=22ltr=22=
><<a href=3D=22mailto:marceloerlemans=40gmail.com=22 target=3D=22=5Fbl=
ank=22>marceloerlemans=40gmail.com</a>></span> wrote:<br><blockquote s=
tyle=3D=22margin:0 0 0 .8ex;border-left:1px =23ccc solid;padding-left:1ex=
=22><div dir=3D=22ltr=22>Hi Christian,<div><br></div><div>Those numbers a=
re impressive=21</div><div>I see you tested default Prevayler up to only =
8 threads. You might want to scale that up a bit=3F </div></div></bl=
ockquote><div><br></div><div>Did this already and it peaked at about 900 =
threads here with 18k operations per second.</div><div> </div><block=
quote style=3D=22margin:0 0 0 .8ex;border-left:1px =23ccc solid;padding-l=
eft:1ex=22><div dir=3D=22ltr=22><div>After I read your mail, I ran a prev=
ayler scalability test, from 1 to 1500 threads, 20 seconds per round/</di=
v><div>In my test setup Prevayler handled 60 transactions/second =40 8 th=
reads, but at 1489 threads, Prevayler could handle 5945 transactions / se=
cond. That's =7E100 times as many. Results in graph form can be found her=
e: <a href=3D=22http://imgur.com/sRIaCVD=22 target=3D=22=5Fblank=22>=
http://imgur.com/sRIaCVD</a></div></div></blockquote><div><br></div><div>=
What machine are you running the test on=3F</div><div> </div><blockq=
uote style=3D=22margin:0 0 0 .8ex;border-left:1px =23ccc solid;padding-le=
ft:1ex=22><div dir=3D=22ltr=22><div>I wanted to do a similar test with Pr=
evayler8, but I see now I need two classes which I cant seem to find:&nbs=
p;ByteBufferOutputStream and ByteBufferInputStream. Can you help me =
with that=3F<br></div><div><br></div></div></blockquote><div><br></div><d=
iv>See the first mail - and mind the purpose and limitation of that imple=
mentation: shines at single-threaded environments (no sync) and sacrifice=
s some Durability points to speed. I'm playing with memory.force() at the=
moment and will integrate that soon.</div><div><br></div><div>What I'm g=
oing to test next, is the default implementation configured for speed: no=
deep copy, no forced disk sync.</div><div><br></div><div>And draw some c=
harts, too. ;-)</div><div><br></div><div>Cheers,</div><div>Christian</div=
></div></div></div>
</div><div><div>---------------------------------------------------------=
---------------------</div><div>Comprehensive Server Monitoring with Site=
24x7.</div><div>Monitor 10 servers for =249/Month.</div><div>Get alerted =
through email, SMS, voice calls or mobile push notifications.</div><div>T=
ake corrective actions from your mobile device.</div><div><a href=3D=22ht=
tp://p.sf.net/sfu/Zoho=22>http://p.sf.net/sfu/Zoho</a></div></div><div><d=
iv>=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=
=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F</di=
v><div>To unsubscribe go to the end of this page: <a href=3D=22http://lis=
ts.sourceforge.net/lists/listinfo/prevayler-discussion=22>http://lists.so=
urceforge.net/lists/listinfo/prevayler-discussion</a></div><div>=5F=5F=5F=
=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=
=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F</div><div>=22Da=
tabases in Memoriam=22 -- <a href=3D=22http://www.prevayler.org=22>http:/=
/www.prevayler.org</a></div></div></div></span>
=20
=20
=20
=20
</blockquote>
=20
<div>
<br>
</div>
--5446c991_dcdf8f6_104--
--===============3913037630437737279==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
------------------------------------------------------------------------------
Comprehensive Server Monitoring with Site24x7.
Monitor 10 servers for $9/Month.
Get alerted through email, SMS, voice calls or mobile push notifications.
Take corrective actions from your mobile device.
http://p.sf.net/sfu/Zoho
--===============3913037630437737279==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
_______________________________________________
To unsubscribe go to the end of this page: http://lists.sourceforge.net/lists/listinfo/prevayler-discussion
_______________________________________________
"Databases in Memoriam" -- http://www.prevayler.org
--===============3913037630437737279==--