No subject
Wed Mar 11 16:46:39 EDT 2009
by loading the exe "text pages" as this can usual be achieve sub
second. The limiting factor seem to be more the loading of data pages
i.e. maps from disk. This purely relies on the OS's disk cache.
I don't deny that on OS's that support the features u point out there
are definite benefits but I'm unsure as to the scale of their impact
on long lived server exe's
I'd be interested to know the results, if I ever get any free time I'd
have a play myself but alas that's never gonna happen :P
Steve / K
P.S. If the above is a bit garbled sorry Christmas + pub and all
that :)
----- Original Message -----=20
From: Mark J. DeFilippis=20
To: cod at icculus.org ; cod at icculus.org=20
Sent: 10 December 2004 13:11
Subject: Re: [cod] Multiple COD:UO servers on the same box
I will have to investigate this further. Which distributions are you referring to?
This would be a crying shame! I would hardly think it would
make sense to change the use of the sticky bit. It's intent
and impact was indeed quite real. This would be more like
MS Office programming, not an O.S. Fat cats get at it?
If every invocation of Linux is copied to swap, they have
regressed to days prior to AT&T Unix System V.3. Why
would they do this? Sell more Ram? Enjoy waiting while
the same text pages from the same executable are loaded
to swap?
On my RH Enterprise 3.0. Fresh boot, loading a utility I
developed a ways back, loads itself, then recursively calls
subroutines, and allocates normal heap and stack (by default,
it can be set, and has some Shared Memory options as well),
from my olden days, shows on repeated load, the amount of
virtual memory utilization did not increase, nor did the swap
space allocated.
(Commercial distributions we run on our IBM platforms, even
up to the time they switched from AIX acted in this manner...
I can't say it is this way for all distributions. However if this is
the case, than IMPO, Linux just became an inferior Cheap
Unix. I wouldn't mind... if it not been removed!
I have seen your writing before, and indeed I believe you, and
the quote, I just "cant' believe" the quote. Are they referring
to perhaps a "regular" non-executable file (I refer to in Kernel
speak as text, hope that did not confuse anyone..)... When I say
"Text pages", I am referring to the "executable code" that is either
static, or dynamic through a shared "reentrant" library....
I will be able to check RH 9.0, Mandrake 10.1, (which I just
received an update on) SuSe 7.3, and Slackware 7.1
This would really bum me out Steve...
=20
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
This e.mail is private and confidential between Multiplay (UK) Ltd. and the person or entity to whom it is addressed. In the event of misdirection, the recipient is prohibited from using, copying, printing or otherwise disseminating it or any information contained in it.=20
In the event of misdirection, illegible or incomplete transmission please telephone (023) 8024 3137
or return the E.mail to postmaster at multiplay.co.uk.
------=_NextPart_000_03D9_01C4DEBC.A096ECC0
Content-Type: text/html;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=3DContent-Type content=3D"text/html; charset=3Diso-8859-1">
<META content=3D"MSHTML 6.00.2900.2523" name=3DGENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=3D#d4d0c8>
<DIV><FONT size=3D2>I here ur there man. The quotes are from the 5.2.1-RELEASE=20
FreeBSD</FONT></DIV>
<DIV><FONT size=3D2>man page. From what I've seen as it doesn't allocate swap for=20
text pages</FONT></DIV>
<DIV><FONT size=3D2>it wouldn't suffer as much from this as an OS that</FONT><FONT=20
size=3D3> does. This can be</FONT></DIV>
<DIV>seen but the fact we run 90% of our machines with just 50MB swap</DIV>
<DIV>and even that never even starts to get allocated. This may well have</DIV>
<DIV>been a deliberate decision with RAM as cheap as it is now :)</DIV>
<DIV> </DIV>
<DIV><FONT size=3D2>From playing with servers day to day it seem they are in no=20
way limited</FONT></DIV>
<DIV><FONT size=3D2>by loading the exe "text pages" as this can usual be achieve=20
sub</FONT></DIV>
<DIV><FONT size=3D2>second. The limiting factor seem to be more the loading of=20
data pages</FONT></DIV>
<DIV><FONT size=3D2>i.e. maps from disk. This purely relies on the OS's disk=20
cache.</FONT></DIV>
<DIV><FONT size=3D2></FONT> </DIV>
<DIV><FONT size=3D2>I don't deny that on OS's that support the features u point=20
out there</FONT></DIV>
<DIV><FONT size=3D2>are definite benefits but I'm unsure as to the scale of their=20
impact</FONT></DIV>
<DIV><FONT size=3D2>on long lived server exe's</FONT></DIV>
<DIV><FONT size=3D2></FONT> </DIV>
<DIV><FONT size=3D2>I'd be interested to know the results, if I ever get any free=20
time I'd</FONT></DIV>
<DIV><FONT size=3D2>have a play myself but alas that's never gonna happen=20
:P</FONT></DIV>
<DIV><FONT size=3D2></FONT> </DIV>
<DIV><FONT size=3D2> Steve / K</FONT></DIV>
<DIV><FONT size=3D2>P.S. If the above is a bit garbled sorry Christmas + pub and=20
all</FONT></DIV>
<DIV><FONT size=3D2>that :)</FONT></DIV>
<DIV><FONT size=3D2></FONT> </DIV>
<DIV><FONT size=3D3>----- Original Message ----- </FONT></DIV>
<BLOCKQUOTE=20
style=3D"PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
<DIV=20
style=3D"BACKGROUND: #e4e4e4; FONT: 10pt arial; font-color: black"><B>From:</B>=20
<A title=3Ddefilm at acm.org href=3D"mailto:defilm at acm.org">Mark J. DeFilippis</A>=20
</DIV>
<DIV style=3D"FONT: 10pt arial"><B>To:</B> <A title=3Dcod at icculus.org=20
href=3D"mailto:cod at icculus.org">cod at icculus.org</A> ; <A title=3Dcod at icculus.org=20
href=3D"mailto:cod at icculus.org">cod at icculus.org</A> </DIV>
<DIV style=3D"FONT: 10pt arial"><B>Sent:</B> 10 December 2004 13:11</DIV>
<DIV style=3D"FONT: 10pt arial"><B>Subject:</B> Re: [cod] Multiple COD:UO=20
servers on the same box</DIV>
<DIV><FONT size=3D2></FONT><BR></DIV><BR>I will have to investigate this=20
further. Which distributions are you referring to?<BR><BR>This would be=20
a crying shame! I would hardly think it would<BR>make sense to change=20
the use of the sticky bit. It's intent<BR>and impact was indeed quite=20
real. This would be more like<BR>MS Office programming, not an O.S. Fat=20
cats get at it?<BR><BR>If every invocation of Linux is copied to swap, they=20
have<BR>regressed to days prior to AT&T Unix System V.3. Why<BR>would they=20
do this? Sell more Ram? Enjoy waiting while<BR>the same text pages from=20
the same executable are loaded<BR>to swap?<BR><BR>On my RH Enterprise=20
3.0. Fresh boot, loading a utility I<BR>developed a ways back, loads=20
itself, then recursively calls<BR>subroutines, and allocates normal heap and=20
stack (by default,<BR>it can be set, and has some Shared Memory options as=20
well),<BR>from my olden days, shows on repeated load, the amount of<BR>virtual=20
memory utilization did not increase, nor did the swap<BR>space=20
allocated.<BR><BR>(Commercial distributions we run on our IBM platforms,=20
even<BR>up to the time they switched from AIX acted in this manner...<BR>I=20
can't say it is this way for all distributions. However if this is<BR>the=20
case, than IMPO, Linux just became an inferior Cheap<BR>Unix. I wouldn't=20
mind... if it not been removed!<BR><BR>I have seen your writing before, and=20
indeed I believe you, and<BR>the quote, I just "cant' believe" the=20
quote. Are they referring<BR>to perhaps a "regular" non-executable file=20
(I refer to in Kernel<BR>speak as text, hope that did not confuse=20
anyone..)... When I say<BR>"Text pages", I am referring to the=20
"executable code" that is either<BR>static, or dynamic through a shared =20
"reentrant" library....<BR><BR>I will be able to check RH 9.0, Mandrake 10.1,=20
(which I just<BR>received an update on) SuSe 7.3, and Slackware 7.1<BR>This=20
would really bum me out=20
Steve...<BR> <BR><TT></BLOCKQUOTE></TT><br>=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D<br>
This e.mail is private and confidential between Multiplay (UK) Ltd. and the person or entity to whom it is addressed. In the event of misdirection, the recipient is prohibited from using, copying, printing or otherwise disseminating it or any information contained in it. <br>
<br>
In the event of misdirection, illegible or incomplete transmission please telephone (023) 8024 3137<br>
or return the E.mail to postmaster at multiplay.co.uk.</body></HTML>
------=_NextPart_000_03D9_01C4DEBC.A096ECC0--
More information about the cod
mailing list