Mersenne Digest       Monday, November 10 2003       Volume 01 : Number 1090




----------------------------------------------------------------------

Date: Mon, 3 Nov 2003 08:12:51 -0800 (PST)
From: Quantum Mechanic <[EMAIL PROTECTED]>
Subject: Mersenne: 50% CPU?

I have an Intel 4 box from MDG, 1 CPU, 2.4GHz, 512MB
RAM, L1 8KB, L2 512KB.

Prime95 is only getting 50%, with System Idle Process
taking 50%. 

It's currently running an LL test in the 20M range.

Any ideas why it's only 50%?

- -QM


=====
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Quantum Mechanics: The dreams stuff is made of

__________________________________
Do you Yahoo!?
Exclusive Video Premiere - Britney Spears
http://launch.yahoo.com/promos/britneyspears/
_________________________________________________________________________
Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm
Mersenne Prime FAQ      -- http://www.tasam.com/~lrwiman/FAQ-mers

------------------------------

Date: Mon, 3 Nov 2003 10:58:33 -0800
From: "Aaron" <[EMAIL PROTECTED]>
Subject: RE: Mersenne: 50% CPU?

Pentium 4's use hyperthreading to give you a 'virtual' second CPU.

Don't bother trying to get another instance of Prime95 running on that 2nd
CPU... It's virtual and while a lot of programs can benefit, the extreme
workout that Prime95 generates means you won't get a benefit from running a
second copy.  It does mean that running a single instance as usual will, in
theory, work even better because a lot more of the OS related things or if
you use the computer to run Word, Excel or whatever will work better because
those can now use the virtual CPU for a lot more things.

And the reason it shows 50% is because Task Manager adds up all the CPU's
and that becomes the 100% high mark.  Since Prime95 is only using the first
CPU, it can at most only use 50%.

If you had multiple *real* processors and had Prime95 running on each one,
it would come out to 100%.

Hope that helps.

> -----Original Message-----
> From: [EMAIL PROTECTED] 
> [mailto:[EMAIL PROTECTED] On Behalf Of 
> Quantum Mechanic
> Sent: Monday, November 03, 2003 8:13 AM
> To: [EMAIL PROTECTED]
> Subject: Mersenne: 50% CPU?
> 
> I have an Intel 4 box from MDG, 1 CPU, 2.4GHz, 512MB RAM, L1 
> 8KB, L2 512KB.
> 
> Prime95 is only getting 50%, with System Idle Process taking 50%. 
> 
> It's currently running an LL test in the 20M range.
> 
> Any ideas why it's only 50%?
> 
> -QM
> 
> 
> =====
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Quantum Mechanics: The dreams stuff is made of
> 
> __________________________________
> Do you Yahoo!?
> Exclusive Video Premiere - Britney Spears 
> http://launch.yahoo.com/promos/britneyspears/
> ______________________________________________________________
> ___________
> Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm
> Mersenne Prime FAQ      -- http://www.tasam.com/~lrwiman/FAQ-mers

_________________________________________________________________________
Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm
Mersenne Prime FAQ      -- http://www.tasam.com/~lrwiman/FAQ-mers

------------------------------

Date: Mon, 3 Nov 2003 14:18:24 EST
From: [EMAIL PROTECTED]
Subject: Re: Mersenne: 50% CPU?

- --part1_7a.4a1fb80c.2cd80400_boundary
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit

In a message dated 11/3/2003 2:04:50 PM Eastern Standard Time, 
[EMAIL PROTECTED] writes:

> Pentium 4's use hyperthreading to give you a 'virtual' second CPU.
> 
> Don't bother trying to get another instance of Prime95 running on that 2nd
> CPU... It's virtual and while a lot of programs can benefit, the extreme
> workout that Prime95 generates means you won't get a benefit from running a
> second copy.  It does mean that running a single instance as usual will, in
> theory, work even better because a lot more of the OS related things or if
> you use the computer to run Word, Excel or whatever will work better because
> those can now use the virtual CPU for a lot more things.
> 
> And the reason it shows 50% is because Task Manager adds up all the CPU's
> and that becomes the 100% high mark.  Since Prime95 is only using the first
> CPU, it can at most only use 50%.
> 
> If you had multiple *real* processors and had Prime95 running on each one,
> it would come out to 100%.
> 
> Hope that helps.
> 
> >-----Original Message-----
> >From: [EMAIL PROTECTED] 
> >[mailto:[EMAIL PROTECTED] On Behalf Of 
> >Quantum Mechanic
> >Sent: Monday, November 03, 2003 8:13 AM
> >To: [EMAIL PROTECTED]
> >Subject: Mersenne: 50% CPU?
> >
> >I have an Intel 4 box from MDG, 1 CPU, 2.4GHz, 512MB RAM, L1 
> >8KB, L2 512KB.
> >
> >Prime95 is only getting 50%, with System Idle Process taking 50%. 
> >
> >It's currently running an LL test in the 20M range.
> >
> >Any ideas why it's only 50%?
> >
> >-QM
> >
> >
> >=====
> >~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >Quantum Mechanics: The dreams stuff is made of
> >
> 

  I have a Pentium 4 processsor and TM indicates an average usage of close to 
100% CPU for Prime 95 - am I running 'multiple instances' without being aware 
of it? If so, how do I prevent this if it will degrade efficiency?

- --part1_7a.4a1fb80c.2cd80400_boundary
Content-Type: text/html; charset="US-ASCII"
Content-Transfer-Encoding: quoted-printable

<HTML><FONT FACE=3Darial,helvetica><BODY BGCOLOR=3D"#ffffff"><FONT  style=
=3D"BACKGROUND-COLOR: #ffffff" SIZE=3D2 FAMILY=3D"SANSSERIF" FACE=3D"Arial"=20=
LANG=3D"0">In a message dated 11/3/2003 2:04:50 PM Eastern Standard Time, aa=
[EMAIL PROTECTED] writes:<BR>
<BR>
<BLOCKQUOTE TYPE=3DCITE style=3D"BORDER-LEFT: #0000ff 2px solid; MARGIN-LEFT=
: 5px; MARGIN-RIGHT: 0px; PADDING-LEFT: 5px">Pentium 4's use hyperthreading=20=
to give you a 'virtual' second CPU.<BR>
<BR>
Don't bother trying to get another instance of Prime95 running on that 2nd<B=
R>
CPU... It's virtual and while a lot of programs can benefit, the extreme<BR>
workout that Prime95 generates means you won't get a benefit from running a<=
BR>
second copy.&nbsp; It does mean that running a single instance as usual will=
, in<BR>
theory, work even better because a lot more of the OS related things or if<B=
R>
you use the computer to run Word, Excel or whatever will work better because=
<BR>
those can now use the virtual CPU for a lot more things.<BR>
<BR>
And the reason it shows 50% is because Task Manager adds up all the CPU's<BR=
>
and that becomes the 100% high mark.&nbsp; Since Prime95 is only using the f=
irst<BR>
CPU, it can at most only use 50%.<BR>
<BR>
If you had multiple *real* processors and had Prime95 running on each one,<B=
R>
it would come out to 100%.<BR>
<BR>
Hope that helps.<BR>
<BR>
&gt;-----Original Message-----<BR>
&gt;From: [EMAIL PROTECTED] <BR>
&gt;[mailto:[EMAIL PROTECTED] On Behalf Of <BR>
&gt;Quantum Mechanic<BR>
&gt;Sent: Monday, November 03, 2003 8:13 AM<BR>
&gt;To: [EMAIL PROTECTED]<BR>
&gt;Subject: Mersenne: 50% CPU?<BR>
&gt;<BR>
&gt;I have an Intel 4 box from MDG, 1 CPU, 2.4GHz, 512MB RAM, L1 <BR>
&gt;8KB, L2 512KB.<BR>
&gt;<BR>
&gt;Prime95 is only getting 50%, with System Idle Process taking 50%. <BR>
&gt;<BR>
&gt;It's currently running an LL test in the 20M range.<BR>
&gt;<BR>
&gt;Any ideas why it's only 50%?<BR>
&gt;<BR>
&gt;-QM<BR>
&gt;<BR>
&gt;<BR>
&gt;=3D=3D=3D=3D=3D<BR>
&gt;~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~<BR>
&gt;Quantum Mechanics: The dreams stuff is made of<BR>
&gt;<BR>
</BLOCKQUOTE><BR>
<BR>
&nbsp; I have a Pentium 4 processsor and TM indicates an average usage of cl=
ose to 100% CPU for Prime 95 - am I running 'multiple instances' without bei=
ng aware of it? If so, how do I prevent this if it will degrade efficiency?<=
/FONT></HTML>

- --part1_7a.4a1fb80c.2cd80400_boundary--
_________________________________________________________________________
Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm
Mersenne Prime FAQ      -- http://www.tasam.com/~lrwiman/FAQ-mers

------------------------------

Date: Mon, 3 Nov 2003 11:39:27 -0800
From: "Aaron" <[EMAIL PROTECTED]>
Subject: RE: Mersenne: 50% CPU?

Does your CPU have hyper-threading for sure, and it's enabled in the BIOS?

Task Manager will show one graph per CPU... If you only see a single graph,
your OS only thinks it has one CPU (counting all virtual and all real CPUs).

I also assume you're using WinNT/Win2k/WinXP since Win98/ME doesn't support
multiple CPU's (and doesn't really have Task Manager either...just that
system resources thing).

Check your list of running processes and see how many Prime95's show up.
Should just be 1.  If you have 2 running, yeah, running 2 processes on a
single CPU, even with the multi-threading, will generally hamper how well
either one runs.

I once toyed with the idea of running LL tests on the real CPU and running
factoring tests on the virtual one, but I never could decide if that was
helping out at all...My gut tells me no just because of the memory accesses
and not using the single L1/L2 caches on the CPU as well as they could be.

By the way, that's a sight to behold when you have a quad CPU P4 system... 8
graphs happily bouncing away.  I'm sure on an 8 CPU system it must be
terribly impressive to have 16 graphs showing up, but I've yet to get my
hands on such a system. :) 

>   I have a Pentium 4 processsor and TM indicates an average 
> usage of close to 100% CPU for Prime 95 - am I running 
> 'multiple instances' without being aware of it? If so, how do 
> I prevent this if it will degrade efficiency?

_________________________________________________________________________
Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm
Mersenne Prime FAQ      -- http://www.tasam.com/~lrwiman/FAQ-mers

------------------------------

Date: Mon, 3 Nov 2003 15:18:21 EST
From: [EMAIL PROTECTED]
Subject: Re: Mersenne: 50% CPU?

- --part1_149.1b68b452.2cd8120d_boundary
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit

In a message dated 11/3/2003 2:43:30 PM Eastern Standard Time, 
[EMAIL PROTECTED] writes:

> Does your CPU have hyper-threading for sure, and it's enabled in the BIOS?
> 
> Task Manager will show one graph per CPU... If you only see a single graph,
> your OS only thinks it has one CPU (counting all virtual and all real CPUs).
> 
> I also assume you're using WinNT/Win2k/WinXP since Win98/ME doesn't support
> multiple CPU's (and doesn't really have Task Manager either...just that
> system resources thing).
> 
> Check your list of running processes and see how many Prime95's show up.
> Should just be 1.  If you have 2 running, yeah, running 2 processes on a
> single CPU, even with the multi-threading, will generally hamper how well
> either one runs.
> 
> I once toyed with the idea of running LL tests on the real CPU and running
> factoring tests on the virtual one, but I never could decide if that was
> helping out at all...My gut tells me no just because of the memory accesses
> and not using the single L1/L2 caches on the CPU as well as they could be.
> 
> By the way, that's a sight to behold when you have a quad CPU P4 system... 8
> graphs happily bouncing away.  I'm sure on an 8 CPU system it must be
> terribly impressive to have 16 graphs showing up, but I've yet to get my
> hands on such a system. :) 
> 

  Thanks very much for the information - my configuration must be to treat 
any virtual CPU as part of the physical one since there is only one graph; 
(unfortunately for the nice visuals - 16 would be lovely and probably hypnotic!).

   I double checked and there is only one instance of Prime 95 running - on 
my older Pentium 3 I had played around with running two copies but the 
interation times were approxiamtely 3 times longer for each than 1 running solo - 
haven't tried it yet on this machine but my impression is the same as yours - 
there wouldn't be much, if any, benefit but it might be fun to play around with 
:-). 

- --part1_149.1b68b452.2cd8120d_boundary
Content-Type: text/html; charset="US-ASCII"
Content-Transfer-Encoding: quoted-printable

<HTML><FONT FACE=3Darial,helvetica><BODY BGCOLOR=3D"#ffffff"><FONT  style=
=3D"BACKGROUND-COLOR: #ffffff" SIZE=3D2 FAMILY=3D"SANSSERIF" FACE=3D"Arial"=20=
LANG=3D"0">In a message dated 11/3/2003 2:43:30 PM Eastern Standard Time, aa=
[EMAIL PROTECTED] writes:<BR>
<BR>
<BLOCKQUOTE TYPE=3DCITE style=3D"BORDER-LEFT: #0000ff 2px solid; MARGIN-LEFT=
: 5px; MARGIN-RIGHT: 0px; PADDING-LEFT: 5px">Does your CPU have hyper-thread=
ing for sure, and it's enabled in the BIOS?<BR>
<BR>
Task Manager will show one graph per CPU... If you only see a single graph,<=
BR>
your OS only thinks it has one CPU (counting all virtual and all real CPUs).=
<BR>
<BR>
I also assume you're using WinNT/Win2k/WinXP since Win98/ME doesn't support<=
BR>
multiple CPU's (and doesn't really have Task Manager either...just that<BR>
system resources thing).<BR>
<BR>
Check your list of running processes and see how many Prime95's show up.<BR>
Should just be 1.&nbsp; If you have 2 running, yeah, running 2 processes on=20=
a<BR>
single CPU, even with the multi-threading, will generally hamper how well<BR=
>
either one runs.<BR>
<BR>
I once toyed with the idea of running LL tests on the real CPU and running<B=
R>
factoring tests on the virtual one, but I never could decide if that was<BR>
helping out at all...My gut tells me no just because of the memory accesses<=
BR>
and not using the single L1/L2 caches on the CPU as well as they could be.<B=
R>
<BR>
By the way, that's a sight to behold when you have a quad CPU P4 system... 8=
<BR>
graphs happily bouncing away.&nbsp; I'm sure on an 8 CPU system it must be<B=
R>
terribly impressive to have 16 graphs showing up, but I've yet to get my<BR>
hands on such a system. :) <BR>
</BLOCKQUOTE><BR>
<BR>
&nbsp; Thanks very much for the information - my configuration must be to tr=
eat any virtual CPU as part of the physical one since there is only one grap=
h; (unfortunately for the nice visuals - 16 would be lovely and probably hyp=
notic!).<BR>
<BR>
&nbsp;&nbsp; I double checked and there is only one instance of Prime 95 run=
ning - on my older Pentium 3 I had played around with running two copies but=
 the interation times were approxiamtely 3 times longer for each than 1 runn=
ing solo - haven't tried it yet on this machine but my impression is the sam=
e as yours - there wouldn't be much, if any, benefit but it might be fun to=20=
play around with :-). </FONT></HTML>

- --part1_149.1b68b452.2cd8120d_boundary--
_________________________________________________________________________
Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm
Mersenne Prime FAQ      -- http://www.tasam.com/~lrwiman/FAQ-mers

------------------------------

Date: Mon, 3 Nov 2003 14:21:38 -0600
From: "Shane Sanford" <[EMAIL PROTECTED]>
Subject: Re: Mersenne: 50% CPU?

This is a multi-part message in MIME format.

- ------=_NextPart_000_000E_01C3A215.CB3E3350
Content-Type: text/plain;
        charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

  I have a Pentium 4 processor and TM indicates an average usage of =
close to 100% CPU for Prime 95 - am I running 'multiple instances' =
without being aware of it? If so, how do I prevent this if it will =
degrade efficiency?=20

=20

Not all P4's have hyper threading & not all motherboards support it on =
top of that.  So it requires a HT enabled CPU, HT enabled Motherboard, & =
HT supported OS for HT to function.  From my experience HT is a nice =
feature since it gives Prime95 a chance to run even when other "not =
nice" programs are running that don't release the processor as often as =
they should even if it's not busy.  The down side is with HT enabled =
Prime95 is more likely to interfere with certain performance sensitive =
applications.

 =20

If you want to learn more about Prime95 & HT there has been several =
recent posts on the topic at

=20

http://www.mersenneforum.org/forumdisplay.php?s=3D&forumid=3D9

=20

There is some evidence that running a LL (high usage of SSE2 floating =
point calculations) & TF (high usage of SSE2 integer calculations) may =
increase overall throughput.

=20

Shane

- ------=_NextPart_000_000E_01C3A215.CB3E3350
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 6.00.2800.1264" name=3DGENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=3D#ffffff>
<DIV><FONT face=3Darial,helvetica><FONT lang=3D0 =
style=3D"BACKGROUND-COLOR: #ffffff"=20
face=3DArial size=3D2 FAMILY=3D"SANSSERIF">
<P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt"><SPAN=20
style=3D"FONT-SIZE: 10pt; BACKGROUND: white; FONT-FAMILY: Arial">&nbsp; =
I have a=20
Pentium 4 processor and TM indicates an average usage of close to 100% =
CPU for=20
Prime 95 - am I running 'multiple instances' without being aware of it? =
If so,=20
how do I prevent this if it will degrade efficiency?</SPAN><FONT =
size=3D3><SPAN=20
style=3D"FONT-FAMILY: Arial"> </SPAN><?xml:namespace prefix =3D o ns =3D =

"urn:schemas-microsoft-com:office:office" /><o:p></o:p></FONT></P>
<P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt"><FONT size=3D3><FONT=20
face=3D"Times New Roman">&nbsp;<o:p></o:p></FONT></FONT></P>
<P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt"><SPAN=20
style=3D"FONT-SIZE: 10pt; FONT-FAMILY: Arial">Not all P4's have hyper =
threading=20
&amp; not all motherboards support it on top of that.&nbsp; So it =
requires a HT=20
enabled CPU, HT enabled Motherboard, &amp; HT supported OS for HT to=20
function.&nbsp; From my experience HT is a nice feature since it gives =
Prime95 a=20
chance to run even when other "not nice" programs are running&nbsp;that =
don't=20
release the processor as often as they should even if it's not =
busy.&nbsp; The=20
down side is with HT enabled Prime95 is more likely to interfere with =
certain=20
performance sensitive applications.</SPAN><o:p></o:p></P>
<P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt"><FONT size=3D3><FONT=20
face=3D"Times New Roman">&nbsp;</FONT></FONT><FONT size=3D3><FONT=20
face=3D"Times New Roman">&nbsp;<o:p></o:p></FONT></FONT></P>
<P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt"><SPAN=20
style=3D"FONT-SIZE: 10pt; FONT-FAMILY: Arial">If you want to learn more =
about=20
Prime95 &amp; HT there has been several recent posts on the topic=20
at</SPAN><o:p></o:p></P>
<P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt"><FONT size=3D3><FONT=20
face=3D"Times New Roman">&nbsp;<o:p></o:p></FONT></FONT></P>
<P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt"><SPAN=20
style=3D"FONT-SIZE: 10pt; FONT-FAMILY: Arial"><A=20
href=3D"http://www.mersenneforum.org/forumdisplay.php?s=3D&amp;forumid=3D=
9">http://www.mersenneforum.org/forumdisplay.php?s=3D&amp;forumid=3D9</A>=
</SPAN><o:p></o:p></P>
<P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt"><FONT size=3D3><FONT=20
face=3D"Times New Roman">&nbsp;<o:p></o:p></FONT></FONT></P>
<P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt"><SPAN=20
style=3D"FONT-SIZE: 10pt; FONT-FAMILY: Arial">There is some evidence =
that running=20
a LL (high usage of SSE2 floating point calculations) &amp; TF (high =
usage of=20
SSE2 integer calculations) may increase overall=20
throughput.</SPAN><o:p></o:p></P>
<P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt"><FONT size=3D3><FONT=20
face=3D"Times New Roman">&nbsp;<o:p></o:p></FONT></FONT></P>
<P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt"><SPAN=20
style=3D"FONT-SIZE: 10pt; FONT-FAMILY: =
Arial">Shane</SPAN><o:p></o:p></P></FONT></FONT></DIV></BODY></HTML>

- ------=_NextPart_000_000E_01C3A215.CB3E3350--


_________________________________________________________________________
Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm
Mersenne Prime FAQ      -- http://www.tasam.com/~lrwiman/FAQ-mers

------------------------------

Date: Mon, 03 Nov 2003 13:23:49 -0800
From: John R Pierce <[EMAIL PROTECTED]>
Subject: Re: Mersenne: 50% CPU?

p4 with 400 or 533Mhz FSB do not support hyperthreading, *EXCEPT* the 
p4-3.06/533 does.

P4 with 800MHz FSB (these are the "C" versions) do support hyperthreading.

P4 Xeons all support HT.



I have a question... gotta couple of dual P4/Xeon 2.8 servers, running 
linux 2.4.20 kernels...   using two instances of mprime on them...   any 
way to insure that the two instances run on actual different CPUs and 
not just two threads of he same CPU?  per 'top`, they seem to have 
landed on CPU2 and CPU3 (out of 0-3).... ooops, now they are on 1,2, 
which means they have NOT got locked CPU affinity.    does anyone know 
which of the 0,1,2,3 CPUs are the physical ones?  meaning, are 0,1 one 
cpu and 2,3 the other, or are 0,2 one and 1,3 the other?



_________________________________________________________________________
Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm
Mersenne Prime FAQ      -- http://www.tasam.com/~lrwiman/FAQ-mers

------------------------------

Date: Mon, 3 Nov 2003 13:34:48 -0800
From: "Aaron" <[EMAIL PROTECTED]>
Subject: RE: Mersenne: 50% CPU?

> Not all P4's have hyper threading & not all motherboards 
> support it on top of that.  So it requires a HT enabled CPU, 
> HT enabled Motherboard, & HT supported OS for HT to function. 

Hmm... Yeah, I thought about that but I guess I just assumed he knew he had
a hyper-threading CPU.

Isn't it just the P4 Xeon's that have that?  Obviously not all P4's have
'em.  I have a P4 laptop and it doesn't have hyper-threading... Wish it did.

The P4 "Extreme Edition" should be shipping now... Has the 2MB L2 cache and
hyperthreading, just like the Xeon's do.

<http://www.intel.com/products/desktop/processors/pentium4HTXE/index.htm?iid
=HPAGE+low_prod_031103&>

Yeah baby... Me wants one of those. :)

PS - Oh... I just checked... I guess "regular" P4's do have hyperthreading
as well, but only in certain models (2.4GHz or faster).  My puny laptop is
only a 1.7GHz P4. :(  No hyperthreading for me.

Here's a link to the Intel hyperthreading page:
<http://www.intel.com/products/desktop/processors/pentium4/index.htm?iid=ipp
_htm+p4p&>

_________________________________________________________________________
Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm
Mersenne Prime FAQ      -- http://www.tasam.com/~lrwiman/FAQ-mers

------------------------------

Date: Mon, 3 Nov 2003 14:45:30 -0800
From: "Aaron" <[EMAIL PROTECTED]>
Subject: RE: Mersenne: 50% CPU?

> p4 with 400 or 533Mhz FSB do not support hyperthreading, *EXCEPT* the
> p4-3.06/533 does.
> 
> P4 with 800MHz FSB (these are the "C" versions) do support 
> hyperthreading.
> 
> P4 Xeons all support HT.

Ahh... Good info.

> does 
> anyone know 
> which of the 0,1,2,3 CPUs are the physical ones?  meaning, 
> are 0,1 one cpu and 2,3 the other, or are 0,2 one and 1,3 the other?

The standard way to identify is to have all physical CPU's listed first (0,1
in a dual CPU system) followed by the virtual CPU's (2,3)

Prime95's INI file lets you set the affinity, so look in the settings of the
program to set that to the physical processors.

I can only vouch for Win32 and how it numbers the processors, but I seem to
recall that Linux numbered them the same way... All physical followed by all
virtual.  I could be wrong though.  I sometimes am. :)

_________________________________________________________________________
Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm
Mersenne Prime FAQ      -- http://www.tasam.com/~lrwiman/FAQ-mers

------------------------------

Date: Mon, 03 Nov 2003 16:35:21 -0800
From: John R Pierce <[EMAIL PROTECTED]>
Subject: Re: Mersenne: 50% CPU?

> The standard way to identify is to have all physical CPU's listed first (0,1
> in a dual CPU system) followed by the virtual CPU's (2,3)
> 

Hmmm, looks like you are wrong...  /proc/cpuinfo shows cpu 0,1 as being 
one physical id, and cpu 2,3 as being the other..

        # cat /proc/cpuinfo
        processor       : 0
        physical id     : 0
        
        processor       : 1
        physical id     : 0
        
        processor       : 2
        physical id     : 3
        
        processor       : 3
        physical id     : 3

(much kruft deleted)

 > Prime95's INI file lets you set the affinity, so look in the settings 
 > of the program to set that to the physical processors.
 >
 > I can only vouch for Win32 and how it numbers the processors, but I
 > seem to recall that Linux numbered them the same way... All physical
 > followed by all virtual.  I could be wrong though.  I sometimes am. :)

I don't know if the affinity stuff will work in linux, however... I 
found some kernel patches that ADD cpu affinity support to linux, but it 
was only folded into the master source as of 2.5.8 (and, presumably, 
2.6.x) and I'm running 2.4.20 on these servers.  I don't really want to 
mess with patching the kernel on these R&D servers, they are being used 
for database development.

meanwhile, watching the CPUs in TOP, the 2 instances are bouncing all 
over them every few seconds.


_________________________________________________________________________
Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm
Mersenne Prime FAQ      -- http://www.tasam.com/~lrwiman/FAQ-mers

------------------------------

Date: Mon, 03 Nov 2003 19:55:10 -0500
From: George Woltman <[EMAIL PROTECTED]>
Subject: Re: Mersenne: 50% CPU?

At 04:35 PM 11/3/2003 -0800, John R Pierce wrote:
>I don't know if the affinity stuff will work in linux, however..

Mprime ignores the affinity settings in the ini files.  We'll have to
add it when linux 2.6 is released. 

_________________________________________________________________________
Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm
Mersenne Prime FAQ      -- http://www.tasam.com/~lrwiman/FAQ-mers

------------------------------

Date: Mon, 03 Nov 2003 17:08:50 -0800
From: John R Pierce <[EMAIL PROTECTED]>
Subject: Re: Mersenne: 50% CPU?

George Woltman wrote:

> At 04:35 PM 11/3/2003 -0800, John R Pierce wrote:
> 
>> I don't know if the affinity stuff will work in linux, however..
> 
> 
> Mprime ignores the affinity settings in the ini files.  We'll have to
> add it when linux 2.6 is released.

kernel 2.6 has been 'released' for some time now.   however, Red Hat 
just announced they are discontinuing all support and future development 
for "Red Hat Linux" effective almost immediately (Dec31 2003 for RH 7.x 
and 8.0, and april 31 2004 for RH 9.0), instead offering Red Hat 
Enterprise 3.0, which costs $$$ per system.  I don't know if Kernel 
2.6.x has made it into any of the other mainstream distros yet.



_________________________________________________________________________
Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm
Mersenne Prime FAQ      -- http://www.tasam.com/~lrwiman/FAQ-mers

------------------------------

Date: Mon, 03 Nov 2003 22:09:25 -0500
From: Nathan Russell <[EMAIL PROTECTED]>
Subject: Re: Mersenne: 50% CPU?

- --On Monday, November 03, 2003 5:08 PM -0800 John R Pierce <[EMAIL PROTECTED]> wrote:

> George Woltman wrote:
>
>> At 04:35 PM 11/3/2003 -0800, John R Pierce wrote:
>>
>>> I don't know if the affinity stuff will work in linux, however..
>>
>>
>> Mprime ignores the affinity settings in the ini files.  We'll have to
>> add it when linux 2.6 is released.
>
> kernel 2.6 has been 'released' for some time now.   however, Red Hat just announced 
> they are discontinuing all support and future development for "Red Hat Linux" 
> effective almost immediately (Dec31 2003 for RH 7.x and 8.0, and april 31 2004 for 
> RH 9.0), instead offering Red Hat Enterprise 3.0, which costs $$$ per system.  I 
> don't know if Kernel 2.6.x has made it into any of the other mainstream distros yet.

Essentially they are discontinuing it under that name, but merging it into the Fedora 
distro (a fully redistributable, open source distro based on RH).  It is simply no 
longer going to be called "Red Hat Linux."  There's some speculation that they may be 
using Fedora for testing code that's not known to be stable before including it in 
their enterprise distro, but that's not the end of the world - anyone can install any 
old rpm version they feel confident in.

See the thread I started at <http://slashdot.org/comments.pl?sid=84550&cid=7379004> 
for more than you wanted to know about the whole thing.

Nathan
_________________________________________________________________________
Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm
Mersenne Prime FAQ      -- http://www.tasam.com/~lrwiman/FAQ-mers

------------------------------

Date: Mon, 10 Nov 2003 07:22:54 -0800 (PST)
From: Chris Marble <[EMAIL PROTECTED]>
Subject: Mersenne: Exponent already tested?

I've been working on a large exponent for a while that I got from Primenet.
A while ago I started getting messages like this:
Sending expected completion date for M33469273: Jan 28 2004
ERROR 11: Exponent already tested.

Does this mean that I was poached?  Will I still get credit if/when I complete
the exponent?
- -- 
  [EMAIL PROTECTED] - HMC UNIX Systems Manager
  My opinions are my own and probably don't represent anything anyway.
_________________________________________________________________________
Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm
Mersenne Prime FAQ      -- http://www.tasam.com/~lrwiman/FAQ-mers

------------------------------

End of Mersenne Digest V1 #1090
*******************************

Reply via email to