CpuIdle 2.5 public beta available

[Andreas Goetz]

FREEWARE

CpuIdle runs a HLT command in an idle priority thread under Win95/98. That
allows modern microprocessors to save power and stay cool. Great for
overclocking.

CpuIdle's is also the best and most complete CPU optimizer for Win95/98 that
exists. Most CPUs include performance increasing options that are disabled
by default. CpuIdle activates them all!

What's new in Version 2.5 (March, 29th, 1998):

- program code completely rewritten
- CpuIdle now completely and automatically optimizes your CPU for
performance enhancement
- removed cpuidle.vxd
- removed -loadonly switch since it didn't offer failure protection
If you need -loadonly try the loadonly.exe standalone program
- added far better power saving for P5 and P5MMX (1.5A..1.8A less)
- added -quit switch to exit CpuIdle after CPU optimization
- removed all Cyrix specific switches from 2.4 (default is now enabled)
- removed -noID switch (nobody used this anyway)
- removed -wait switch (no longer needed)
- added new switches to selectively disable optimization
- workaround for Cyrix and IDT clock rate bug
- added CPUID for Cyrix CPUs even if disabled
- DOS based optimizers like set6x86 no longer needed for complete Cyrix
support
- added new direct95.vxd driver to get access to

Also included in the release archive is -= DosIdle 2.00 =- by Marton Balog
Used in conjunction with CpuIdle, DosIdle and CpuIdle provide a complete
power saving and optimization solution both for Win95/98 and DOS, even
incompatible DOS programs in Win95 DOS boxes are supported. Please notice
that DosIdle is not a CpuIdle addon but a standalone program. These programs
are best used together and packaged for greater convenience.

Andreas Goetz
mailto:go...@stud.uni-hannover.de
http://www.stud.uni-hannover.de/~goetz
A Sysops' Work Is Never Done.

[KALLE]

I' using an old version of it and it is great!

KALLE

see http://www.jump.net/~lcs/kalle.htm for ASUS P55T2P4 information.

[Okeedokee]

Previous versions work great. 2.5 gives me the blue screen of death.
Direct95.VXD error.
Andreas Goetz wrote in message
<6f83rs$sud$1...@newsserver.rrzn.uni-hannover.de>...

[Shawn]

I get the same error .vxd error. No blue screen though. Previous versions
worked fine. I also notice what appears to be the fans spinning faster
creating noise. It goes away after 5-10min. I have popped the hood to
check it out yet. I like the added features though.

-Shawn

Okeedokee wrote in message <35192...@ntnews.compusmart.ab.ca>...

[Jadawin]

Shawn schrieb in Nachricht <6fbdfp$4o$1...@news.ncal.verio.com>...

>I get the same error .vxd error. No blue screen though. Previous versions
>worked fine. I also notice what appears to be the fans spinning faster
>creating noise. It goes away after 5-10min. I have popped the hood to
>check it out yet. I like the added features though.
>
>-Shawn

I get no errors... everything works fine.

Jadawin

[Bryan Rozon]

I also get the Blue Screen.

[Steve & Dandi]

Perhaps you and Bryan Rozon and Shawn should mention which cpu you are
using to see if there is something in common?

Okeedokee wrote:
>
> Previous versions work great. 2.5 gives me the blue screen of death.
> Direct95.VXD error.
>
> Andreas Goetz wrote in message
> <6f83rs$sud$1...@newsserver.rrzn.uni-hannover.de>...
> >FREEWARE
> >

[Okeedokee]

Downloaded 2.5b2 with the fix and it works great.

[Seamuso]

I've downloaded and installed with no problems what so ever. TX97XE with AMD K6
233. Avereage system temp with no heavy load is at 35 degrees.

[Shawn]

Steve & Dandi wrote in message <3519BE...@mindspring.com>...

>Perhaps you and Bryan Rozon and Shawn should mention which cpu you are
>using to see if there is something in common?
>
>Okeedokee wrote:
>>
>> Previous versions work great. 2.5 gives me the blue screen of death.
>> Direct95.VXD error.

I downloaded 2.5b2 and everything seems ok now. Thanks for the quick fix!
I am running a P100 and a P150.

Where can I find a list of the optimizations?

Thanks again for a great program.

- Shawn

[Philip]

I can now up the new version 2.5b2. With this version the cpu temperature
is dropped by 4 degrees C. Fantastic!
Bryan Rozon wrote in message <35195...@news.cadvision.com>...

>I also get the Blue Screen.
>
>Okeedokee wrote in message <35192...@ntnews.compusmart.ab.ca>...

>>Previous versions work great. 2.5 gives me the blue screen of death.
>>Direct95.VXD error.

[Craig Matchan]

Hi.

Okeedokee wrote:

Previous versions work great. 2.5 gives me the blue screen of death.
Direct95.VXD error.

The 1st beta of 2.5 had a bug that caused it to crash on  some CPUs. To quote from his page....

Start quote
 "Beta 1 contained a very serious bug that caused GPFs on P5, P5MMX and C6 chips (the curse
 of cut and paste). Beta 2 should fix this. If you experience problems please run with "-noopt" switch,
 go to the Cpu Info sheet, copy all infos (especially CPU type and stepping and send me a detailed bug
 report -- Thanks)."
End quote.

CPUIdle 2.5B2 is now available for downloading which is meant to fix the problems.

Craig
--
-----------------------------------------------------------------
Craig Matchan                             Open Learning Australia
Systems Administrator                    6/30 Colins St Melbourne
                                                  Australia, 3000
Voice: 61 3 9903 8948             email: craig....@ola.edu.au
Fax:   61 3 9903 8976   www: http://gromit.adm.monash.edu.au/~chm
-----------------------------------------------------------------
 

[Bryan Rozon]

I am using P55 166 MMX.

Steve & Dandi wrote in message <3519BE...@mindspring.com>...
>Perhaps you and Bryan Rozon and Shawn should mention which cpu you are
>using to see if there is something in common?
>

>Okeedokee wrote:
>>
>> Previous versions work great. 2.5 gives me the blue screen of death.
>> Direct95.VXD error.
>>

>> Andreas Goetz wrote in message
>> <6f83rs$sud$1...@newsserver.rrzn.uni-hannover.de>...
>> >FREEWARE
>> >

[Stephen Urgo]

I have same cpu overclocked to 200mhz, and get same error using 2.5.
Previous versions of cpuidle worked fine.
My system:

Asus tx97e motherboard
64 sdram
pent 166mmx oc to 200mhz
Quantum fireball st 3.2 hd
Win98 beta 3 (4.10.1650) with IE4.01 on clean HD

Any suggestions?

Steve

Browsing With KeyBored
(Disability and Computers - maybe there IS a connection...)
http://www.geocities.com/Heartland/Ranch/4039/

Bryan Rozon wrote in message <3519f...@news.cadvision.com>...

>I am using P55 166 MMX.

[Douglas J. Young]

In article <3519e...@ntnews.compusmart.ab.ca>, Okeedokee
<pha...@ontis.com> writes

>Downloaded 2.5b2 with the fix and it works great.
>
>

Anyone know of anyway to measure the temp of the processor via software
(I have onboard thermal measuring but it only says if it is above 55c or
not, which it isn't thankfully)>
--
Douglas J. Young

Hellfire Warrior "Burnt" Level 50
Sorcerer "Hellfire" Level 50

Diablo Warrior "Burnt" Level 50
Sorcerer "Hellfire" Level 50

The Hellfire Basement/ZImage Graphics :-
Http://www.troopers.demon.co.uk

(Only working with Internet Explorer at the moment, Netscape in a couple of
weeks).
---
Hitting her with a frying pan is a good idea. We might want to have a backup
plan though, just in case she isn't a cartoon.
---
So I figure I'll be crazy man with a snake, you know, crazy snake man. Then I'll
get more snakes, call them my babies, kids will walk past my place they will run
"Run away from crazy snake man" they'll shout!

[Gary Fox]

Go back and get the update to 2.5.

Terry/Sonny wrote in message <351af2b...@news.tm.net.my>...
>I get a blue screen everytime i try to run cpuidle 2.5. it says
>something about direct95.vxd or something like that.
>
>-Terry
>

[Stephanie Lynn]

On Tue, 24 Mar 1998 10:24:03 +0100, "Andreas Goetz"
<go...@stud.uni-hannover.de> wrote:

Does anyone know if it will run on a P-II 333Mhz???

System:
Intel P-II 333Mhz
128MB SDRam
8.4GB IDE HD
3.5GB IDE HD
1.6GB IDE HD
32x IDE CD-ROM
JAZ External SCSI Drive

Any information that you can provide on the support of the P-II Chip
would be nice.

Stephanie

[Terry/Sonny]

[Wayne Ernst]

Microsoft ought to reimburse everyone $20 who
purchases this product. It's their OS that is not
operating as it should with the 'halt' instruction.
The person who created this utility could retire
at a very young age.

Dizway <Dizway@[see.sig]> wrote in article
<351af813...@news.dizway.com>...
> Gonna send him his 20 bucks a.s.a.p. Great product. Thanx man
>
> Dizway
>
> --
> D i z w a y @ n y m . a l i a s . n e t
> Email will be PGP encrypted to me.
> Best viewed w/ Forte Agent - Courier +10
>

[Jim]

2.5 Beta 3 is out

On Tue, 24 Mar 1998 10:24:03 +0100, "Andreas Goetz"
<go...@stud.uni-hannover.de> wrote:

>FREEWARE
>
>CpuIdle runs a HLT command in an idle priority thread under Win95/98. That
>allows modern microprocessors to save power and stay cool. Great for
>overclocking.
>

[oacs]

Whaaaat? Ive got Beta 2, and thats really great. My CPU now cools down 4 deg
more than with ver2.3!!! Seems quite aggressive!

In article <6ff4br$f...@bgtnsc02.worldnet.att.net>,

[df]

Where might I obtain this utility? Please reply to...
vide...@bcn.net

Philip wrote in message <6fbqt0$mvh$1...@newsource.ihug.co.nz>...

>I can now up the new version 2.5b2. With this version the cpu temperature
>is dropped by 4 degrees C. Fantastic!
>Bryan Rozon wrote in message <35195...@news.cadvision.com>...
>>I also get the Blue Screen.
>>
>>Okeedokee wrote in message <35192...@ntnews.compusmart.ab.ca>...

>>>Previous versions work great. 2.5 gives me the blue screen of death.
>>>Direct95.VXD error.

[p971...@unitele.com.my]

"Philip" <wk...@ihug.co.nz> wrote:
> I can now up the new version 2.5b2. With this version the cpu temperature
> is dropped by 4 degrees C. Fantastic!

How do you measure the temperature ?

-----== Posted via Deja News, The Leader in Internet Discussion ==-----
http://www.dejanews.com/ Now offering spam-free web-based newsreading

[Jim Dompier]

In article <RbwIWBAk...@troopers.demon.co.uk>,
do...@troopers.demon.co.uk says...

> In article <3519e...@ntnews.compusmart.ab.ca>, Okeedokee
> <pha...@ontis.com> writes
> >Downloaded 2.5b2 with the fix and it works great.
> >
> >
> Anyone know of anyway to measure the temp of the processor via software
> (I have onboard thermal measuring but it only says if it is above 55c or
> not, which it isn't thankfully)>
> --
> Douglas J. Young

Try Mother board monitor...

http://www.euronet.nl/users/darkside/mbmonitor/

--
Jim Dompier

**remove the seven and the obvious to reply**

[Jim Dompier]

In article <6fi1ll$bb8$1...@nnrp1.dejanews.com>, p971...@unitele.com.my
says...

Motherboard Montior..

[Wayne Ernst]

Try:
http://www.stud.uni-hannover.de/~goetz/

df <vide...@bcn.net> wrote in article <6fhlsm$igr$1...@news.berk.net>...

> Where might I obtain this utility? Please reply to...
> vide...@bcn.net
>
>
>
>
> Philip wrote in message <6fbqt0$mvh$1...@newsource.ihug.co.nz>...

> >I can now up the new version 2.5b2. With this version the cpu
temperature
> >is dropped by 4 degrees C. Fantastic!

[wayne smith]

what is the downside of using cpuidle? does the cpu run slower ,say
because of the time it takes to power up to speed? and what does this
thing do exactly ,I mean how do you "idle" a cpu?, also I tried cpumon
and the cpu runs at 34.5C and mb at 33C this seams a bit high to me..its
an amd k6 200 o/c to 225 on an asus tx97e at 75hrz.......any comments or
help on this would be great......Wayne

[Steve & Dandi]

Douglas J. Young wrote:
>
> ---
> Hitting her with a frying pan is a good idea. We might want to have a backup
> plan though, just in case she isn't a cartoon.
> ---
> So I figure I'll be crazy man with a snake, you know, crazy snake man. Then I'll
> get more snakes, call them my babies, kids will walk past my place they will run
> "Run away from crazy snake man" they'll shout!

Can you say "Fruitloops"? I new you could.

[ript]

>> >
>what is the downside of using cpuidle? does the cpu run slower ,say
>because of the time it takes to power up to speed? and what does this
>thing do exactly ,I mean how do you "idle" a cpu?, also I tried cpumon
>and the cpu runs at 34.5C and mb at 33C this seams a bit high to me..its
>an amd k6 200 o/c to 225 on an asus tx97e at 75hrz.......any comments or
>help on this would be great......Wayne

it may be like 1% slower on a benchmark.....

[William Mitchell]

On Sat, 28 Mar 1998 20:00:12 GMT, sl...@nospam.yermom.org (ript)
wrote:

With CPUIDLE 2.5 installed my cpu runs 150rpm slower but my cpu temp.
has decreased 6.5C and the MB temp. has decreased 5C.
- Intel P166MMX on a TX97-E running at 75MHz.

Check out the Helix website:
http://www.escape.ca/~helix

Bill Mitchell
Helix Computer Services
Winnipeg, MB
phone: (204) 792-3141
fax: (204) 889-7187
email: he...@escape.ca

[Sean]

Excuse me, but what kind of CPU do you have that runs at REVOLUTION's Per
MINUTE???? That is pretty unique. LOL

I have used CPU idle with a K6200 and let me tell you it works! I had a
very professional thermal probe attached to the cpu and as soon as CPU idle
was engaged - the temperature would instantly begin to drop. It would
always lower the CPU temp by 9-11 degrees celcius. The only thing it seems
to do is cut down on the harddrive Cached speed. Uncached speeds seem to be
unaffected. All applications are uneffected. Basically you cannot tell it
is there.

William Mitchell wrote in message <351d5f3f...@news.escape.ca>...

[ript]

In article <351d5f3f...@news.escape.ca>, he...@escape.ca wrote:
>On Sat, 28 Mar 1998 20:00:12 GMT, sl...@nospam.yermom.org (ript)
>wrote:
>
>>>> >
>>>what is the downside of using cpuidle? does the cpu run slower ,say
>>>because of the time it takes to power up to speed? and what does this
>>>thing do exactly ,I mean how do you "idle" a cpu?, also I tried cpumon
>>>and the cpu runs at 34.5C and mb at 33C this seams a bit high to me..its
>>>an amd k6 200 o/c to 225 on an asus tx97e at 75hrz.......any comments or
>>>help on this would be great......Wayne
>>
>>it may be like 1% slower on a benchmark.....
>>
>
>With CPUIDLE 2.5 installed my cpu runs 150rpm slower but my cpu temp.
>has decreased 6.5C and the MB temp. has decreased 5C.
>- Intel P166MMX on a TX97-E running at 75MHz.
>
>
>
>

your cpu spins???

[LShaping]

> >With CPUIDLE 2.5 installed my cpu runs 150rpm slower but my cpu temp.
> >has decreased 6.5C and the MB temp. has decreased 5C.
> >- Intel P166MMX on a TX97-E running at 75MHz.

> your cpu spins???

Maybe it's programmed in Turbo C :)

[LShaping]

> >With CPUIDLE 2.5 installed my cpu runs 150rpm slower but my cpu temp.
> >has decreased 6.5C and the MB temp. has decreased 5C.
> >- Intel P166MMX on a TX97-E running at 75MHz.

> your cpu spins???

At the page http://www.stud.uni-hannover.de/~goetz/
"If you want to know how much a difference CpuIdle makes on your system and
you have a mainboard with the TX chipset you may want to try CpuMonitor
written by Martin Grote which displays mainboard and CPU temperature as
well as fan speed."
He is referring to fan speed.

[Adam Stouffer]

Stephen Urgo wrote:
>
> Win98 beta 3 (4.10.1650) with IE4.01 on clean HD
>
> Any suggestions?
>

fdisk that mofo and run linux. CPU's will always run cooler
with linux.
--
_________________________

|
|
| Remove the NO-SPAM |
|
|
_________________________

[William Mitchell]

On Sat, 28 Mar 1998 16:07:09 -0500, "Sean" <sean%mcd@logic%south.com>
wrote:

>Excuse me, but what kind of CPU do you have that runs at REVOLUTION's Per
>MINUTE???? That is pretty unique. LOL

Hehehe...reference to the CPU fan. ;)

>
>I have used CPU idle with a K6200 and let me tell you it works! I had a
>very professional thermal probe attached to the cpu and as soon as CPU idle
>was engaged - the temperature would instantly begin to drop. It would
>always lower the CPU temp by 9-11 degrees celcius. The only thing it seems
>to do is cut down on the harddrive Cached speed. Uncached speeds seem to be
>unaffected. All applications are uneffected. Basically you cannot tell it
>is there.
>
>
>William Mitchell wrote in message <351d5f3f...@news.escape.ca>...

>>On Sat, 28 Mar 1998 20:00:12 GMT, sl...@nospam.yermom.org (ript)
>>wrote:
>>
>>>>> >
>>>>what is the downside of using cpuidle? does the cpu run slower ,say
>>>>because of the time it takes to power up to speed? and what does this
>>>>thing do exactly ,I mean how do you "idle" a cpu?, also I tried cpumon
>>>>and the cpu runs at 34.5C and mb at 33C this seams a bit high to me..its
>>>>an amd k6 200 o/c to 225 on an asus tx97e at 75hrz.......any comments or
>>>>help on this would be great......Wayne
>>>
>>>it may be like 1% slower on a benchmark.....
>>>
>>

>>With CPUIDLE 2.5 installed my cpu runs 150rpm slower but my cpu temp.
>>has decreased 6.5C and the MB temp. has decreased 5C.
>>- Intel P166MMX on a TX97-E running at 75MHz.
>>
>>
>>
>>
>>
>>

[LShaping]

> CpuIdle runs a HLT command in an idle priority thread under Win95/98.
That
> allows modern microprocessors to save power and stay cool. Great for
> overclocking.

I like the idea of running my CPU cooler but aren't you being a little
cryptic? Is "an HLT command" an assembler language pseudonym for
something? Can you spell it out? Otherwise <sigh> your wording is nearly
meaningless to me.

[Seamuso]

Anyone have any success running Cpuidle and MotherBoard Monitor from linux using
a windows emulator?

SeamusO

[Tom Brennan]

I don't program in assembler these days, but the old 8080 and Z80 chips
had a halt instruction (I'm sure that must be what HLT is here) that
would stop processing at that point and wait for an interrupt before
continuing. I seem to remember an 808x chip where the halt would cause
the external buss to go 'tri-state' (neither a 1 nor a zero) which could
cut the current use down a bit. I would guess today's chips have all
sorts of things they can shut down internally to save a few watts while
waiting for work to do.

Tom Brennan

[Gary Fox]

How about Halt (HLT), as in stop processing, turn off, don't run, stay cool!

LShaping wrote in message <6fkmi7$d...@bgtnsc02.worldnet.att.net>...

[Harry McGregor]

You DO NOT need CPU IDLE in Linux or in NT, both OSes properly
implement the HALT instruction, the same thing that CPU IDLE does. MS
decided that Windows 95 did not need it, and they have it running a
plain idle loop when it is not doing anything. On my Linux box
(486DX33 notebook), I get GREAT battery life since it stops the chip
when it is not needed, and cooling my K6 166@208 WinNT box is not much
of a problem since it also executes the HALT instruction (I do have 8
fans, and a MONSTER heat sink and fan in it).

Harry

[Jim Dompier]

In article <6fj7f2$d...@nntp.logicsouth.com>, sean%mcd@logic%south.com
says...

> Excuse me, but what kind of CPU do you have that runs at REVOLUTION's Per
> MINUTE???? That is pretty unique. LOL
>

He meant Horse Power (HP)...<G>!

[r...@geocities.com]

This software sounds interesting? Is it worth trying?

Anyone know if it works well on an FX chipset? I will try anything to
speed this Packard Bell computer up.

Later
Ray
--
Visit Ray's Packard Bell Web Site
http://www.geocities.com/SiliconValley/Lakes/8774/

[John D. Goulden]

Hi, Ray -- you will not see a performance gain from CPU Idle (in fact, some
users report significantly slower hard drive I/O with CPU Idle). If it
works on your system, your CPU will run a little cooler.

> This software sounds interesting? Is it worth trying?
>
> Anyone know if it works well on an FX chipset? I will try anything to
> speed this Packard Bell computer up.

--
John D. Goulden
Veritas Computing Systems and Services
jgou...@flash.net

[Jim A.]

On Sat, 28 Mar 1998 16:07:09 -0500, "Sean" <sean%mcd@logic%south.com>
wrote:

>Excuse me, but what kind of CPU do you have that runs at REVOLUTION's Per

>MINUTE???? That is pretty unique. LOL
>

My K6-200 burns oil, I have to add a quart every Sunday when I fill up
with unleaded premium.

Other than that it can burn rubber even on the 4th gear.

[Jim A.]

On Sun, 29 Mar 1998 02:43:26 GMT, Adam Stouffer
<te...@NO-SPAM.sgi.net> wrote:

>
>fdisk that mofo and run linux. CPU's will always run cooler
>with linux.

Better yet, sell the computer and get a typewriter, that will run even
cooler.

Jim

[Jim A.]

On Sun, 29 Mar 1998 19:20:42 -0500, r...@geocities.com wrote:

>I will try anything to
>speed this Packard Bell computer up.
>

Attach PB on solid fuel rocket, point towards the sun, ignite!

[LShaping]

> > > CpuIdle runs a HLT command in an idle priority thread under Win95/98.
> > That
> > > allows modern microprocessors to save power and stay cool. Great for
> > > overclocking.

> > I like the idea of running my CPU cooler but aren't you being a little
> > cryptic? Is "an HLT command" an assembler language pseudonym for
> > something? Can you spell it out? Otherwise <sigh> your wording is
nearly
> > meaningless to me.

> I don't program in assembler these days, but the old 8080 and Z80 chips
> had a halt instruction (I'm sure that must be what HLT is here) that
> would stop processing at that point and wait for an interrupt before
> continuing. I seem to remember an 808x chip where the halt would cause
> the external buss to go 'tri-state' (neither a 1 nor a zero) which could
> cut the current use down a bit. I would guess today's chips have all
> sorts of things they can shut down internally to save a few watts while
> waiting for work to do.
>
> Tom Brennan

Tri-state outputs allow for a one, a zero, and a high impedance state.
When the output is high impedance, it looks like an open circuit, like it
isn't even there. This probably would save power in certain situations.
As far as I know, with tri-state outputs, you can connect several outputs
together. If you want to "hear" from only one of them, you set all the
others to high impedance (I think the abbreviation is "X").
Yup, a lot of the new computer chips have "standby" mode. Also, a great
effort is being made to consume less and less current when the device is in
use. The trade off, on any given day, is performance. Parts that use less
electricity don't perform as well or as fast.
I wish my motherboard were state of the art, with built in temperature
sensors, fan speed controllers, ect. I ran the CPUIdle program. It
reduced performance a little. It causes Windows 95 "System Monitor" to
show CPU activity to be %100 all of the time. CPUs typically get replaced
before they wear out, so as long as your CPU does not exceed
specifications, I wouldn't worry about it. However, having a temperature
sensor/alarm, and maybe to have an automatic shutdown at a too high
temperature is probably the better solution.
LShaping.

[Gary]

On 30 Mar 1998 06:04:11 GMT, "LShaping" <LSha...@worldnet.att.net>
wrote:

If that were true Vacuum tubes and/or discrete transistors would be
the *ultimate* choices for a processor due to their high
power/electricity consumption.

>I wish my motherboard were state of the art, with built in temperature
>sensors, fan speed controllers, ect. I ran the CPUIdle program. It
>reduced performance a little. It causes Windows 95 "System Monitor" to
>show CPU activity to be %100 all of the time.

No doubt a *bug* with the way System Monitor reports activity because
that's exactly the opposite of what's happening. CPUIdle is stopping
most all of the heat generating full on/off clock cycles during times
when there's no activity required from the processor. In other words
why have the car engine (CPU) turning at 3000 rpm when you're waiting
at a red light?

> CPUs typically get replaced
>before they wear out, so as long as your CPU does not exceed
>specifications, I wouldn't worry about it. However, having a temperature
>sensor/alarm, and maybe to have an automatic shutdown at a too high
>temperature is probably the better solution.
>LShaping.

With today's speedy current hungry processors and case ventilation
that *may* leave something to be desired I'd recommend giving it a try
if one has questions about the temperature of their CPU. Unlike Temp
sensors etc. CPU Idle is *free* (unless you feel you'd like to donate)
and the same sort of process is *included* in NT etc. so it's not like
it's some sort of untested theory.

Gary ....

[ript]

In article <351f4910...@news.usol.com>, gd...@rocketmail.com (Gary) wrote:
<<<snip>>>

>
>No doubt a *bug* with the way System Monitor reports activity because
>that's exactly the opposite of what's happening. CPUIdle is stopping
>most all of the heat generating full on/off clock cycles during times
>when there's no activity required from the processor. In other words
>why have the car engine (CPU) turning at 3000 rpm when you're waiting
>at a red light?

<<<<snip>>>>

most processor mesuring utilities actualy maesure time spent in the idle loop.
when running cpuidle itisnt being used as the cpu is in HLT when it would
normanly be running idle.....

[Robert Fries]

sl...@nospam.yermom.org (ript) wrote:

Yes - CPUIDLE will show up as 100% CPU - but it doesn't matter - it's at a low
priority, and instantly gets out of the way. BTW, if you subject the system
to FULL load, meaning there's always a process ready to run on the queue,
the CPU will get just as hot as without CPUIDLE, so you still should have good
cooling if you overclock.

/RF

Robert Fries
rfries at netcom dot com

[seamuso]

WOOHOO....K6 233 oc success 250mhz (3x83)...cpuidle keeping temps at 35 degrees
(average), no more than 55 when running intensive progs

SeamusO

[Robert Fries]

I think you're agreeing with me. Just remember, if your system is
truly busy, CPUIDLE won't keep it any cooler. By 'truly busy' I mean
there's ALWAYS something ready to run.

I'm curious - how are you people measuring CPU temp? It seems to me
that it would be very sesitive to the exact location of the probe.

/RF

[Jeff Gentry]

Robert Fries (rfr...@concentric.net) wrote:
: I think you're agreeing with me. Just remember, if your system is

: truly busy, CPUIDLE won't keep it any cooler. By 'truly busy' I mean
: there's ALWAYS something ready to run.

Exactly. I'm often running a particular app that consumes 100% of free
CPU time on my K6-200. It is not uncommon for me to leave this on
20 hours a day. When i'm running it, my CPU is hot as ever. When i'm not,
my cpu is ice cold (using the "hand test") ... the first time i felt it,
I said, "damn hot!" but it was so cold that i thought it was hot.

--
Jeff Gentry jes...@rpi.edu RPI CompSci Senior
SEX DRUGS UNIX

[Byron Bodo]

In article <35212EB7...@concentric.net>, rfr...@concentric.net says...

>
>I think you're agreeing with me. Just remember, if your system is
>truly busy, CPUIDLE won't keep it any cooler. By 'truly busy' I mean
>there's ALWAYS something ready to run.
>

>I'm curious - how are you people measuring CPU temp? It seems to me
>that it would be very sesitive to the exact location of the probe.

Likely.

Is it feasible to measure power consumption as an alternative
measure of CPUidle's effectiveness, particularly for K6s?

-bb

[Buckaroo]

Adam Stouffer wrote in message <351DB6...@NO-SPAM.sgi.net>...

>
>fdisk that mofo and run linux. CPU's will always run cooler
>with linux.

No argument here. CPU ain't doing anything during a "kernel panic".

[LShaping]

> >I'm curious - how are you people measuring CPU temp? It seems to me
> >that it would be very sesitive to the exact location of the probe.

> Is it feasible to measure power consumption as an alternative
> measure of CPUidle's effectiveness, particularly for K6s?

The newer motherboards have temp sensors built in.

[LShaping]

> >..... The trade off, on any given day, is performance. Parts that

> >use less electricity don't perform as well or as fast.
>
> If that were true Vacuum tubes and/or discrete transistors would be
> the *ultimate* choices for a processor due to their high
> power/electricity consumption.

Wrong. You did not understand what I said. Note that I said "on any given
day". That means that whatever the state of the art happens to be ON ANY
GIVEN DAY, less power means less performance. Besides, I was not bashing
your program. I was discussing electronics hardware :)

> >I wish my motherboard were state of the art, with built in temperature
> >sensors, fan speed controllers, ect. I ran the CPUIdle program. It
> >reduced performance a little. It causes Windows 95 "System Monitor" to
> >show CPU activity to be %100 all of the time.
>

> No doubt a *bug* with the way System Monitor reports activity because
> that's exactly the opposite of what's happening.

System Monitor works just fine without "CpuIdle" installed.

> > CPUs typically get replaced
> >before they wear out, so as long as your CPU does not exceed
> >specifications, I wouldn't worry about it. However, having a
temperature
> >sensor/alarm, and maybe to have an automatic shutdown at a too high
> >temperature is probably the better solution.
> >LShaping.
>
> With today's speedy current hungry processors and case ventilation
> that *may* leave something to be desired I'd recommend giving it a try
> if one has questions about the temperature of their CPU.

> Gary.

Noooo! If one has questions about the temperature of their CPU, answer the
questions first!
Good luck :)
LShaping.

[LShaping]

> I think you're agreeing with me. Just remember, if your system is
> truly busy, CPUIDLE won't keep it any cooler. By 'truly busy' I mean
> there's ALWAYS something ready to run.

> I'm curious - how are you people measuring CPU temp? It seems to me
> that it would be very sesitive to the exact location of the probe.

> /RF

I have installed a temperature sensor on a PC. I used a thermister. A
thermister changes resistance as temperature changes. I put it between the
CPU and the motherboard, touching the underside (what is probably a ground
plane) of the CPU. Temperature sensors are about the size of a BB and cost
about fifty cents, see National Semiconductor's website, search for LM58.
Also, turning the fan off when the system is idle is simple. This matter
is similar to the old PCs that had no real time clock. New motherboards
have these features.
LShaping.

[Byron Bodo]

In article <6fs3cr$g...@bgtnsc02.worldnet.att.net>, LSha...@worldnet.att.net says...

>
>> >I'm curious - how are you people measuring CPU temp? It seems to me
>> >that it would be very sesitive to the exact location of the probe.
>
>

>> Is it feasible to measure power consumption as an alternative
>> measure of CPUidle's effectiveness, particularly for K6s?

>The newer motherboards have temp sensors built in.

Nonetheless, any significant reduction in power consumption
would be worth noting. Heat = energy so there has to
some link.

-bb

[Gary]

On 1 Apr 1998 01:24:02 GMT, "LShaping" <LSha...@worldnet.att.net>
wrote:

>> >..... The trade off, on any given day, is performance. Parts that
>> >use less electricity don't perform as well or as fast.
>>
>> If that were true Vacuum tubes and/or discrete transistors would be
>> the *ultimate* choices for a processor due to their high
>> power/electricity consumption.
>
>Wrong. You did not understand what I said. Note that I said "on any given
>day". That means that whatever the state of the art happens to be ON ANY
>GIVEN DAY, less power means less performance. Besides, I was not bashing
>your program. I was discussing electronics hardware :)

Nope, I don't see any law of physics that would support that statement
as far as switching states or speed is concerned. My K6-233 runs at a
3.2 or 3.3 volt core voltage. some people have *REDUCED* that to 3.0V
or 2.8 V to reduce the power consumption/heat. How does that *lessen*
its performance ?? It's still changing states 233 million times per
second. (233 MHz) and processing data as fast as my K6 at 3.3 volts.
Dropping the voltage on a CPU *may* lessen its *reliabilty* but it
doesn't lessen its speed/performance.

>
>
>
>
>> >I wish my motherboard were state of the art, with built in temperature
>> >sensors, fan speed controllers, ect. I ran the CPUIdle program. It
>> >reduced performance a little. It causes Windows 95 "System Monitor" to
>> >show CPU activity to be %100 all of the time.
>>
>> No doubt a *bug* with the way System Monitor reports activity because
>> that's exactly the opposite of what's happening.
>
>System Monitor works just fine without "CpuIdle" installed.
>
>
>
>
>> > CPUs typically get replaced
>> >before they wear out, so as long as your CPU does not exceed
>> >specifications, I wouldn't worry about it. However, having a
>temperature
>> >sensor/alarm, and maybe to have an automatic shutdown at a too high
>> >temperature is probably the better solution.
>> >LShaping.
>>
>> With today's speedy current hungry processors and case ventilation
>> that *may* leave something to be desired I'd recommend giving it a try
>> if one has questions about the temperature of their CPU.
>> Gary.
>
>Noooo! If one has questions about the temperature of their CPU, answer the
>questions first!

Any valid reasons ??? the same *process* is built *INTO* NT (the
later versions anyway)

Gary

>Good luck :)
>LShaping.

[Anton Smith]

Excuse me, there is a fair bit of difference between voltage and current.
Voltage is the amount of energy given to each packet of electrons (one
coloumb) and current is the number of coloumbs passing through any given
point at any given time. However, the two are co-related and have a directly
proportional relationship. I believe that CpuIdle reduces the current as it
makes references to the ampage used etc. (amp is the unit for current).
-Anton.

>Nope, I don't see any law of physics that would support that statement
>as far as switching states or speed is concerned. My K6-233 runs at a
>3.2 or 3.3 volt core voltage. some people have *REDUCED* that to 3.0V
>or 2.8 V to reduce the power consumption/heat. How does that *lessen*
>its performance ?? It's still changing states 233 million times per
>second. (233 MHz) and processing data as fast as my K6 at 3.3 volts.
>Dropping the voltage on a CPU *may* lessen its *reliabilty* but it
>doesn't lessen its speed/performance.
>
>

>Gary

[LShaping]

> >> >..... The trade off, on any given day, is performance. Parts that
> >> >use less electricity don't perform as well or as fast.
> >>
> >> If that were true Vacuum tubes and/or discrete transistors would be
> >> the *ultimate* choices for a processor due to their high
> >> power/electricity consumption.
> >
> >Wrong. You did not understand what I said. Note that I said "on any
given
> >day". That means that whatever the state of the art happens to be ON
ANY
> >GIVEN DAY, less power means less performance. Besides, I was not
bashing
> >your program. I was discussing electronics hardware :)
>

> Nope, I don't see any law of physics that would support that statement
> as far as switching states or speed is concerned.

What is "switching states"?

[LShaping]

> >> >I'm curious - how are you people measuring CPU temp? It seems to me
> >> >that it would be very sesitive to the exact location of the probe.
> >
> >
> >> Is it feasible to measure power consumption as an alternative
> >> measure of CPUidle's effectiveness, particularly for K6s?
>
> >The newer motherboards have temp sensors built in.
>
> Nonetheless, any significant reduction in power consumption
> would be worth noting. Heat = energy so there has to
> some link.
> -bb

As long as resistance remains about the same, you can calculate power by
current consumption. So measuring current would do.

[Doug]

gd...@rocketmail.com (Gary) wrote:

>Nope, I don't see any law of physics that would support that statement

>as far as switching states or speed is concerned. My K6-233 runs at a
>3.2 or 3.3 volt core voltage. some people have *REDUCED* that to 3.0V
>or 2.8 V to reduce the power consumption/heat. How does that *lessen*
>its performance ?? It's still changing states 233 million times per
>second. (233 MHz) and processing data as fast as my K6 at 3.3 volts.
>Dropping the voltage on a CPU *may* lessen its *reliabilty* but it
>doesn't lessen its speed/performance.

Sorry, but LShaping is correct. Transistors switch faster when they
have more voltage to work with. With a higher voltage, the
transistor's "stored charge" drains-off faster (more current). More
voltage, more current, more power, more heat, faster switching, higher
performance (assuming of course that you raise your clock to
take-advantage of the higher performance potential).

[LShaping]

> >> >..... The trade off, on any given day, is performance. Parts that
> >> >use less electricity don't perform as well or as fast.
> >>
> >> If that were true Vacuum tubes and/or discrete transistors would be
> >> the *ultimate* choices for a processor due to their high
> >> power/electricity consumption.
> >
> >Wrong. You did not understand what I said. Note that I said "on
> >any given >day". That means that whatever the state of the art
> >happens to be ON ANY >GIVEN DAY, less power means less performance.
> >Besides, I was not bashing >your program. I was discussing
> >electronics hardware :)
>

> Nope, I don't see any law of physics that would support that statement
> as far as switching states or speed is concerned. My K6-233 runs at a
> 3.2 or 3.3 volt core voltage. some people have *REDUCED* that to 3.0V
> or 2.8 V to reduce the power consumption/heat. How does that *lessen*
> its performance ?? It's still changing states 233 million times per
> second. (233 MHz) and processing data as fast as my K6 at 3.3 volts.
> Dropping the voltage on a CPU *may* lessen its *reliabilty* but it
> doesn't lessen its speed/performance.

You are playing around with the hazardous idea that adjusting CPU voltage
is OK. Nonsense. I hope that you are not the same person who wrote the
CpuIdle program because you know nothing about hardware. But I am not in
the right group to continue trying to burn your garbage, so I will leave it
at this.

[Gary]

On Wed, 1 Apr 1998 22:59:34 +1200, "Anton Smith" <a.s...@xtra.co.nz>
wrote:

>Excuse me, there is a fair bit of difference between voltage and current.
>Voltage is the amount of energy given to each packet of electrons (one
>coloumb) and current is the number of coloumbs passing through any given
>point at any given time. However, the two are co-related and have a directly
>proportional relationship. I believe that CpuIdle reduces the current as it
>makes references to the ampage used etc. (amp is the unit for current).
>-Anton.

You've taken the wrong spin on the argument, I was responding to the
previous posters statement that any device that has lower voltage has
less perfomance than one with higher voltage. CPUIdle has *nothing* to
do with voltage or current other than the fact that by it stopping the
clock cycles between interrupts the amount of power *USED* is reduced
, that's why there's less *HEAT* produced. CPUIdle is like a car
engine that goes back to a slow idle speed at a stop light (between
interupts it basically does nothing) thus reducing the amount of power
used.

Gary.

>
>>Nope, I don't see any law of physics that would support that statement
>>as far as switching states or speed is concerned. My K6-233 runs at a
>>3.2 or 3.3 volt core voltage. some people have *REDUCED* that to 3.0V
>>or 2.8 V to reduce the power consumption/heat. How does that *lessen*
>>its performance ?? It's still changing states 233 million times per
>>second. (233 MHz) and processing data as fast as my K6 at 3.3 volts.
>>Dropping the voltage on a CPU *may* lessen its *reliabilty* but it
>>doesn't lessen its speed/performance.
>>
>>

>>Gary
>
>

[Gary]

On 1 Apr 1998 13:18:38 GMT, "LShaping" <LSha...@worldnet.att.net>
wrote:

>> >> >..... The trade off, on any given day, is performance. Parts that
>> >> >use less electricity don't perform as well or as fast.
>> >>
>> >> If that were true Vacuum tubes and/or discrete transistors would be
>> >> the *ultimate* choices for a processor due to their high
>> >> power/electricity consumption.
>> >
>> >Wrong. You did not understand what I said. Note that I said "on any
>given
>> >day". That means that whatever the state of the art happens to be ON
>ANY
>> >GIVEN DAY, less power means less performance. Besides, I was not
>bashing
>> >your program. I was discussing electronics hardware :)
>>

>> Nope, I don't see any law of physics that would support that statement
>> as far as switching states or speed is concerned.
>

>What is "switching states"?

It's what CPUIdle prevents between interrupts , in a normal processor
say a 233 Mhz Pentium the transistors *constantly* switch states from
on to off 233 million times a second whether it's processing data or
not. CPUIdle stops (issues the halt command) the needless switching
of states (that uses power) when there's no data to be processed thus
it saves power when there's nothing for the processor to do (no data
to process) and that's why there's *less* heat. In a normal system the
differnece in heat can be quite dramatic , however if a system is
running an application that requires constant or near constant use of
the processor then the heat saving won't be as great or non-existant.

It can be compared somewhat to a car engine, why have the car engine
run at 3000 RPM at a stoplight (no data to process) when it'll
function just fine at idle speed (shut off in our CPU'S case) and then
speed back up when the light changes (an interrupt has occured - data
to process in our case CPU back actively switching)

Gary

[Stuart]

LShaping wrote:
>
> > >> >..... The trade off, on any given day, is performance. Parts that
> > >> >use less electricity don't perform as well or as fast.
> > >>
> > >> If that were true Vacuum tubes and/or discrete transistors would be
> > >> the *ultimate* choices for a processor due to their high
> > >> power/electricity consumption.
> > >
> > >Wrong. You did not understand what I said. Note that I said "on
> > >any given >day". That means that whatever the state of the art
> > >happens to be ON ANY >GIVEN DAY, less power means less performance.
> > >Besides, I was not bashing >your program. I was discussing
> > >electronics hardware :)
> >
> > Nope, I don't see any law of physics that would support that statement

> > as far as switching states or speed is concerned. My K6-233 runs at a
> > 3.2 or 3.3 volt core voltage. some people have *REDUCED* that to 3.0V
> > or 2.8 V to reduce the power consumption/heat. How does that *lessen*
> > its performance ?? It's still changing states 233 million times per
> > second. (233 MHz) and processing data as fast as my K6 at 3.3 volts.
> > Dropping the voltage on a CPU *may* lessen its *reliabilty* but it
> > doesn't lessen its speed/performance.
>

> You are playing around with the hazardous idea that adjusting CPU voltage
> is OK. Nonsense. I hope that you are not the same person who wrote the
> CpuIdle program because you know nothing about hardware. But I am not in
> the right group to continue trying to burn your garbage, so I will leave it
> at this.
>

Thank God. Maybe you'll quit offering us the "benefit" of your
"expertise".

[Gary]

On Wed, 01 Apr 1998 13:59:16 GMT, do...@dgii.com (Doug) wrote:

>gd...@rocketmail.com (Gary) wrote:
>
>>Nope, I don't see any law of physics that would support that statement
>>as far as switching states or speed is concerned. My K6-233 runs at a
>>3.2 or 3.3 volt core voltage. some people have *REDUCED* that to 3.0V
>>or 2.8 V to reduce the power consumption/heat. How does that *lessen*
>>its performance ?? It's still changing states 233 million times per
>>second. (233 MHz) and processing data as fast as my K6 at 3.3 volts.
>>Dropping the voltage on a CPU *may* lessen its *reliabilty* but it
>>doesn't lessen its speed/performance.
>

>Sorry, but LShaping is correct. Transistors switch faster when they
>have more voltage to work with. With a higher voltage, the
>transistor's "stored charge" drains-off faster (more current).

You need some *basic* electronics courses , tranisistors don't "Store
Charges" they either switch states when used as digital switching
devices (in our uses for them in CPU's Etc.) from one potential to
another, cutoff (No output) = 0 , saturation (full on) = 1 ( That's
where the binary notation of 0's and 1's come in) or else in analog
applications ( amplification) transistors turn on (usually linearlly)
to the amount of base current applied , that's where you see the term
"Linear Amplifier" Used in Stereos, radios and other analog devices.

What you basically described in your above statement is a
"Capcacitor" which holds or stores a charge and not a transistor.

> More
>voltage, more current, more power, more heat, faster switching, higher
>performance (assuming of course that you raise your clock to
>take-advantage of the higher performance potential).

Nonsense, ( you need to study a bit more electronics)

Because of inadvertant capacitance in electronic devices (and perhaps
other factors) and maybe in the circuits themselves there is a period
of time between switching states called the "Slew Rate" (though not
nearly as pronounced it's like a car that goes from 0 to 60 , it
doesn't get to 60 instantaneously) and all other factors being equal
the slew rate is less (less time) to switch between 0 and 3 volts than
there is to switch between 0 and 5 volts or 0 and 100 volts
(difference in the time for a car to go from 0 to 20 VS 0 to 100). All
factors being equal the exact *opposite* of what you said is true , it
takes *less* time to switch between 0 volts and a lower voltage than 0
volts and a higher voltage.

Gary

[Gary]

On 1 Apr 1998 14:27:42 GMT, "LShaping" <LSha...@worldnet.att.net>
wrote:

>> >> >..... The trade off, on any given day, is performance. Parts that
>> >> >use less electricity don't perform as well or as fast.
>> >>
>> >> If that were true Vacuum tubes and/or discrete transistors would be
>> >> the *ultimate* choices for a processor due to their high
>> >> power/electricity consumption.
>> >
>> >Wrong. You did not understand what I said. Note that I said "on
>> >any given >day". That means that whatever the state of the art
>> >happens to be ON ANY >GIVEN DAY, less power means less performance.
>> >Besides, I was not bashing >your program. I was discussing
>> >electronics hardware :)
>>

>> Nope, I don't see any law of physics that would support that statement
>> as far as switching states or speed is concerned. My K6-233 runs at a
>> 3.2 or 3.3 volt core voltage. some people have *REDUCED* that to 3.0V
>> or 2.8 V to reduce the power consumption/heat. How does that *lessen*
>> its performance ?? It's still changing states 233 million times per
>> second. (233 MHz) and processing data as fast as my K6 at 3.3 volts.
>> Dropping the voltage on a CPU *may* lessen its *reliabilty* but it
>> doesn't lessen its speed/performance.
>

>You are playing around with the hazardous idea that adjusting CPU voltage
>is OK. Nonsense. I hope that you are not the same person who wrote the
>CpuIdle program because you know nothing about hardware. But I am not in
>the right group to continue trying to burn your garbage, so I will leave it
>at this.

I never made any such assertion, I merely pointed out that *some*
individuals had lowered the voltage on their CPU's and it's
performance was the same as mine. I was just replying that contrarry
to your statement devices with/using lower voltages do not
neccessarily have lower performance.

Gary

[Paul A. Jacobi]

Byron Bodo wrote:
> Is it feasible to measure power consumption as an alternative
> measure of CPUidle's effectiveness, particularly for K6s?
>

I have a digital watt meter that measures current drawn from the A/C
outlet.
On a Digital Celebris GL6200 (Pentium Pro 256K), I measured that CPUIDLE
drops
power consumption approximately 5 watts. This isn't a very big %
savings in
power, which is probably why Microsoft didn't bother to add support for
this
feature. In contrast, an energy saving monitor can save at least 2 or 3
time
that amount.

+---------------------------------------------------------------------------+
| Paul A. Jacobi Phone: (603)
884-1948 |
| Digital Equipment Corporation FAX : (603)
884-0189 |
| OpenVMS Systems Group, ZKO3-4/U14 Email:
jac...@star.enet.dec.com |
| 110 Spitbrook
Road |
| Nashua, NH
03062-2698 |
|
|
| Anti-spam enabled! To reply, remove "nospam-" from the return
address. |
+---------------------------------------------------------------------------+

[LShaping]

> Thank God. Maybe you'll quit offering us the "benefit" of your
> "expertise".

If you don't like it, put me in your kill file. You do know how to use a
kill file, don't you?

[KALLE]

HI ALL

It would take a lot of time to comment and correct all the wrong
statements of this tread. Maybe I spend some time tomorow.

The fact, that no 'regular poster' took part in this tread speaks for
itself.

KALLE

see http://www.jump.net/~lcs/kalle.htm for ASUS P55T2P4 information.

[LShaping]

> >Nope, I don't see any law of physics that would support that statement
> >as far as switching states or speed is concerned. My K6-233 runs at a
> >3.2 or 3.3 volt core voltage. some people have *REDUCED* that to 3.0V
> >or 2.8 V to reduce the power consumption/heat. How does that *lessen*
> >its performance ?? It's still changing states 233 million times per
> >second. (233 MHz) and processing data as fast as my K6 at 3.3 volts.
> >Dropping the voltage on a CPU *may* lessen its *reliabilty* but it
> >doesn't lessen its speed/performance.
>

> Sorry, but LShaping is correct. Transistors switch faster when they
> have more voltage to work with. With a higher voltage, the

> transistor's "stored charge" drains-off faster (more current). More

> voltage, more current, more power, more heat, faster switching, higher
> performance (assuming of course that you raise your clock to
> take-advantage of the higher performance potential).

He is upset with himself for buying what was a state of the art PC before
they were advanced enough to include simple temperature sensors and fan
speed controllers. Now, he has a fast CPU on a dumb motherboard. My
advice to anybody in the same situation would be to consider upgrading your
motherboard. You can find much better ones for a little over one hundred
dollars.

[KALLE]

On 1 Apr 1998 13:30:47 GMT, "LShaping" <LSha...@worldnet.att.net>
wrote:

>> >> >I'm curious - how are you people measuring CPU temp? It seems to me
>> >> >that it would be very sesitive to the exact location of the probe.
>> >
>> >

>> >> Is it feasible to measure power consumption as an alternative
>> >> measure of CPUidle's effectiveness, particularly for K6s?
>>

>> >The newer motherboards have temp sensors built in.
>>
>> Nonetheless, any significant reduction in power consumption
>> would be worth noting. Heat = energy so there has to
>> some link.
>> -bb
>
>As long as resistance remains about the same, you can calculate power by
>current consumption. So measuring current would do.
>

Hmmmm

- The formula to bring voltage, resistance and current together is:
U = R * I
So if the resistance remains about the same and the current is
changing the voltage must change. too.

And I thought they extra builded Voltage regulators into the Board to
avoid changes in the voltage.

Maybe I misunderstood something?

I thought the power will be calculated P=U*I. The voltage is constant
(?) and so the current is directly proportional to the power
consumption / heat production. (The resistance is changing now, but
you can forget about resistance for this calculation!)

But maybe you bring some new interesting ideas to science.

KALLE

P.S. I'm stupid! I just took a look at the calendar! You made an
April-Joke to us! HA HA HA

[LShaping]

> >> >> >I'm curious - how are you people measuring CPU temp? It seems
> >> >> >to me that it would be very sesitive to the exact location of
> >> >> >the probe.

> >> >> Is it feasible to measure power consumption as an alternative
> >> >> measure of CPUidle's effectiveness, particularly for K6s?

(stuff clipped)

> >As long as resistance remains about the same, you can calculate power
> >by current consumption. So measuring current would do.

> - The formula to bring voltage, resistance and current together is:

> U = R * I
> So if the resistance remains about the same and the current is
> changing the voltage must change. too.
> And I thought they extra builded Voltage regulators into the Board to
> avoid changes in the voltage.
> Maybe I misunderstood something?
> I thought the power will be calculated P=U*I. The voltage is constant
> (?) and so the current is directly proportional to the power
> consumption / heat production. (The resistance is changing now, but
> you can forget about resistance for this calculation!)

Right. The R would change. As long as the regulator is functioning
properly, current would denote power and might predict heat.

> But maybe you bring some new interesting ideas to science.
> KALLE

I try :)

[Jeff Cook]

Andreas, this is pretty cool stuff (no pun intended.)
So... isthere a CPUIDLE or analogue for Linux and/or OS/2?

[Steve & Dandi]

LShaping wrote:
>
> > Nope, I don't see any law of physics that would support that statement
> > as far as switching states or speed is concerned.
>

> What is "switching states"?

The Milwaukee Braves moving to Atlanta. :->

[Steve & Dandi]

Robert Fries wrote:
>
> I think you're agreeing with me. Just remember, if your system is
> truly busy, CPUIDLE won't keep it any cooler. By 'truly busy' I mean
> there's ALWAYS something ready to run.
>

> I'm curious - how are you people measuring CPU temp? It seems to me
> that it would be very sesitive to the exact location of the probe.
>

> /RF

I can't find the article, but someone suggested a Radio Shack Indoor-
Outdoor thermometer with the outdoor probe stuck in the CPU heatsink
fins. I just went to get one, and they have a Clearance (at least, the
store I went to) on cat. 63-1011 ($9.97), which also has an overtemp
alarm. Not too shabby.

I've had it running for an hour and without Cpuidle I get 38C. When I
enabled Cpuidle it dropped 10C within a couple of minutes. Most
excellent.

[LShaping]

> It would take a lot of time to comment and correct all the wrong
> statements of this tread. Maybe I spend some time tomorow.
> The fact, that no 'regular poster' took part in this tread speaks for
> itself.
> KALLE

I would welcome ALL of your corrections. At your convenience, please do.
LShaping.

[ript]

In article <3522ac4f...@news.owl-online.de>, michael.RE...@MEowl-online.de (KALLE) wrote:
>On 1 Apr 1998 13:30:47 GMT, "LShaping" <LSha...@worldnet.att.net>
>wrote:
>

>>> >> >I'm curious - how are you people measuring CPU temp? It seems to me
>>> >> >that it would be very sesitive to the exact location of the probe.
>>> >
>>> >

>>> >> Is it feasible to measure power consumption as an alternative
>>> >> measure of CPUidle's effectiveness, particularly for K6s?
>>>

>>> >The newer motherboards have temp sensors built in.
>>>
>>> Nonetheless, any significant reduction in power consumption
>>> would be worth noting. Heat = energy so there has to
>>> some link.
>>> -bb
>>

>>As long as resistance remains about the same, you can calculate power by
>>current consumption. So measuring current would do.
>>
>

>Hmmmm

>
>- The formula to bring voltage, resistance and current together is:
> U = R * I
>So if the resistance remains about the same and the current is
>changing the voltage must change. too.
>
>And I thought they extra builded Voltage regulators into the Board to
>avoid changes in the voltage.
>
>Maybe I misunderstood something?
>
>I thought the power will be calculated P=U*I. The voltage is constant
>(?) and so the current is directly proportional to the power
>consumption / heat production. (The resistance is changing now, but
>you can forget about resistance for this calculation!)
>

>But maybe you bring some new interesting ideas to science.
>
>KALLE
>

>P.S. I'm stupid! I just took a look at the calendar! You made an
> April-Joke to us! HA HA HA
>
>see http://www.jump.net/~lcs/kalle.htm for ASUS P55T2P4 information.

the voltage doesnt change must at all, its regulated therfore if the current
chatges, then the load impeadence is changing inversly with the current.....

[ript]

In article <352247a6...@news.dgii.com>, do...@dgii.com (Doug) wrote:

>gd...@rocketmail.com (Gary) wrote:
>
>>Nope, I don't see any law of physics that would support that statement

>>as far as switching states or speed is concerned. My K6-233 runs at a
>>3.2 or 3.3 volt core voltage. some people have *REDUCED* that to 3.0V
>>or 2.8 V to reduce the power consumption/heat. How does that *lessen*
>>its performance ?? It's still changing states 233 million times per
>>second. (233 MHz) and processing data as fast as my K6 at 3.3 volts.
>>Dropping the voltage on a CPU *may* lessen its *reliabilty* but it
>>doesn't lessen its speed/performance.
>
>Sorry, but LShaping is correct. Transistors switch faster when they
>have more voltage to work with. With a higher voltage, the
>transistor's "stored charge" drains-off faster (more current). More
>voltage, more current, more power, more heat, faster switching, higher
>performance (assuming of course that you raise your clock to
>take-advantage of the higher performance potential).
>

the higher vlotage alows a specific processor to be clocked higher, yea I'll
buy that. but thats totaly backwards between processor familys......

[ript]

In article <6fuekl$v43$1...@lead.zk3.dec.com>, "Paul A. Jacobi" <jac...@star.enet.dec.nospam.com> wrote:

>Byron Bodo wrote:
>> Is it feasible to measure power consumption as an alternative
>> measure of CPUidle's effectiveness, particularly for K6s?
>>
>

>I have a digital watt meter that measures current drawn from the A/C
>outlet.
>On a Digital Celebris GL6200 (Pentium Pro 256K), I measured that CPUIDLE
>drops
>power consumption approximately 5 watts. This isn't a very big %
>savings in
>power, which is probably why Microsoft didn't bother to add support for
>this
>feature. In contrast, an energy saving monitor can save at least 2 or 3
>time
>that amount.

what is significant is the temp of the CPU......

[ript]

In article <35248d1c...@news.usol.com>, gd...@rocketmail.com (Gary) wrote:

>On Wed, 01 Apr 1998 13:59:16 GMT, do...@dgii.com (Doug) wrote:
>
>>gd...@rocketmail.com (Gary) wrote:
>>
>>>Nope, I don't see any law of physics that would support that statement
>>>as far as switching states or speed is concerned. My K6-233 runs at a
>>>3.2 or 3.3 volt core voltage. some people have *REDUCED* that to 3.0V
>>>or 2.8 V to reduce the power consumption/heat. How does that *lessen*
>>>its performance ?? It's still changing states 233 million times per
>>>second. (233 MHz) and processing data as fast as my K6 at 3.3 volts.
>>>Dropping the voltage on a CPU *may* lessen its *reliabilty* but it
>>>doesn't lessen its speed/performance.
>>
>>Sorry, but LShaping is correct. Transistors switch faster when they
>>have more voltage to work with. With a higher voltage, the
>>transistor's "stored charge" drains-off faster (more current).
>

>You need some *basic* electronics courses , tranisistors don't "Store
>Charges" they either switch states when used as digital switching
>devices (in our uses for them in CPU's Etc.) from one potential to
>another, cutoff (No output) = 0 , saturation (full on) = 1 ( That's
>where the binary notation of 0's and 1's come in) or else in analog
>applications ( amplification) transistors turn on (usually linearlly)
>to the amount of base current applied , that's where you see the term
>"Linear Amplifier" Used in Stereos, radios and other analog devices.
>
> What you basically described in your above statement is a
>"Capcacitor" which holds or stores a charge and not a transistor.

the junction of a transitor has a bit of capacitance, enough to cause problems
at realy high speeds, and it is one ot the MAJOR reasons why processors
disipate more power at higher clock speeds.....

[ript]

In article <6ftisu$f...@bgtnsc03.worldnet.att.net>, "LShaping" <LSha...@worldnet.att.net> wrote:
>> >> >..... The trade off, on any given day, is performance. Parts that
>> >> >use less electricity don't perform as well or as fast.
>> >>
>> >> If that were true Vacuum tubes and/or discrete transistors would be
>> >> the *ultimate* choices for a processor due to their high
>> >> power/electricity consumption.
>> >
>> >Wrong. You did not understand what I said. Note that I said "on
>> >any given >day". That means that whatever the state of the art
>> >happens to be ON ANY >GIVEN DAY, less power means less performance.
>> >Besides, I was not bashing >your program. I was discussing
>> >electronics hardware :)
>>

>> Nope, I don't see any law of physics that would support that statement
>> as far as switching states or speed is concerned. My K6-233 runs at a
>> 3.2 or 3.3 volt core voltage. some people have *REDUCED* that to 3.0V
>> or 2.8 V to reduce the power consumption/heat. How does that *lessen*
>> its performance ?? It's still changing states 233 million times per
>> second. (233 MHz) and processing data as fast as my K6 at 3.3 volts.
>> Dropping the voltage on a CPU *may* lessen its *reliabilty* but it
>> doesn't lessen its speed/performance.
>

>You are playing around with the hazardous idea that adjusting CPU voltage
>is OK. Nonsense. I hope that you are not the same person who wrote the
>CpuIdle program because you know nothing about hardware. But I am not in
>the right group to continue trying to burn your garbage, so I will leave it
>at this.
>

my k6-166 runs at 233 with 3.3 volts on the core and I/O .....

[LShaping]

(much stuff clipped)

> my k6-166 runs at 233 with 3.3 volts on the core and I/O .....

Are you saying that the "I/O" voltage is adjusted separately? My MB has
only one adjustment to 3.45, 4, or 5 volts. I have had trouble
overclocking so I will give another visit to AMDs web site looking for the
CPU data sheet and see if maybe I have room to increase the voltage. But I
wouldn't think so since standard high speed logic (like HCT) has a small
range of operating voltage.

[Barney Google]

On Wed, 01 Apr 1998 23:29:41 GMT, michael.RE...@MEowl-online.de
(KALLE) wrote:

>
>Hmmmm
>
>- The formula to bring voltage, resistance and current together is:
> U = R * I

E = R * I here in the US.

>So if the resistance remains about the same and the current is
>changing the voltage must change. too.
>
>And I thought they extra builded Voltage regulators into the Board to
>avoid changes in the voltage.
>
>Maybe I misunderstood something?

For all practical purposes resistance always remains the same unless
oxidation sets in. Voltage will always remain the same because of the
regulators. What is left is current which is always changing. There is
thing called "duty cycle" which can vary anywhere from 0 to 100%. The
more "on/off" transitions of all the millions of transistors per
second in the CPU the more power is consumed. So if the CPU was told
to go to sleep instead of twiddling it's thumbs the power consumption
would decrease. Which means less heat.

>
>I thought the power will be calculated P=U*I. The voltage is constant

No it's P = E * I. Europeons for some reason are backasswards from
Americans. Or is it the other way around?

[Gary]

On Wed, 01 Apr 1998 17:13:32 -0500, "Paul A. Jacobi"
<jac...@star.enet.dec.nospam.com> wrote:

>Byron Bodo wrote:
>> Is it feasible to measure power consumption as an alternative
>> measure of CPUidle's effectiveness, particularly for K6s?
>>
>
>I have a digital watt meter that measures current drawn from the A/C
>outlet.
>On a Digital Celebris GL6200 (Pentium Pro 256K), I measured that CPUIDLE
>drops
>power consumption approximately 5 watts. This isn't a very big %
>savings in
>power,

CPUIdle wasn't designed to save energy, it was designed to reduce
*heat* given off by lots of current fast processors. In many instances
(depending on the application) people that formerly had to use a huge
heatsink and/or fan or keep their case cover off to keep the CPU from
overheating can now use a normal fan/heatsink and keep the case cover
on. CPUIdle just stops a number of clock cycles between interrupts,
basically when there's no data to process (between interrupts) the
processor is sleeping (very little power drawn) and thus less *heat*

> which is probably why Microsoft didn't bother to add support for
>this

Microsoft *DID* provide the same function, it's in Windows NT but not
in Win95 or Win98

Gary

>feature. In contrast, an energy saving monitor can save at least 2 or 3
>time
>that amount.
>
>

[LShaping]

> >Hmmmm
> >
> >- The formula to bring voltage, resistance and current together is:
> > U = R * I
>
> E = R * I here in the US.

Who noticed?

> >So if the resistance remains about the same and the current is
> >changing the voltage must change. too.
> >
> >And I thought they extra builded Voltage regulators into the Board to
> >avoid changes in the voltage.
> >
> >Maybe I misunderstood something?

> For all practical purposes resistance always remains the same unless
> oxidation sets in. Voltage will always remain the same because of the
> regulators. What is left is current which is always changing. There is
> thing called "duty cycle" which can vary anywhere from 0 to 100%. The
> more "on/off" transitions of all the millions of transistors per
> second in the CPU the more power is consumed. So if the CPU was told
> to go to sleep instead of twiddling it's thumbs the power consumption
> would decrease. Which means less heat.

KALLE is correct. Resistance constantly varies. With a given voltage,
that is what causes change in current. In this context, duty cycle should
always be fifty percent. But you are right about putting the CPU to sleep
or at least reducing the clock speed. That does mean less heat. But even
more important than putting the CPU to sleep is telling the temperature of
your CPU. The program can't help if your CPU is running at full capacity.
Therefore, if your processor is not properely cooled in the first place, it
will self destruct. If you can find a way, you should first make sure that
you can tell if it overheats. Then mess around with freeware.

> >I thought the power will be calculated P=U*I. The voltage is constant
>
> No it's P = E * I. Europeons for some reason are backasswards from
> Americans. Or is it the other way around?

Who noticed?

(stuff clipped)

[Gary]

On Wed, 01 Apr 1998 18:01:34 -0800, Steve & Dandi
<sas...@mindspring.com> wrote:

>Robert Fries wrote:
>>
>> I think you're agreeing with me. Just remember, if your system is
>> truly busy, CPUIDLE won't keep it any cooler. By 'truly busy' I mean
>> there's ALWAYS something ready to run.
>>

>> I'm curious - how are you people measuring CPU temp? It seems to me
>> that it would be very sesitive to the exact location of the probe.
>>

>> /RF
>
>I can't find the article, but someone suggested a Radio Shack Indoor-
>Outdoor thermometer with the outdoor probe stuck in the CPU heatsink
>fins. I just went to get one, and they have a Clearance (at least, the
>store I went to) on cat. 63-1011 ($9.97), which also has an overtemp
>alarm. Not too shabby.
>
>I've had it running for an hour and without Cpuidle I get 38C. When I
>enabled Cpuidle it dropped 10C within a couple of minutes. Most
>excellent.

That was my posting/suggestion for measuring the CPU temp. My results
are almost the exact same as yours. I've found you can disable CPUIdle
and in just a few seconds you can see the temp begin to rise, amazing.

Gary

[LShaping]

(stuff clipped)

> I can't find the article, but someone suggested a Radio Shack Indoor-
> Outdoor thermometer with the outdoor probe stuck in the CPU heatsink
> fins. I just went to get one, and they have a Clearance (at least, the
> store I went to) on cat. 63-1011 ($9.97), which also has an overtemp
> alarm. Not too shabby.

GREAT ADVICE if the alarm setting can be high enough to allow for operation
at full capacity. Sounds like the ideal solution for those of you who have
expensive CPUs sitting on outdated motherboards.
What is the maximum temperature it will measure?

[Gary]

On Thu, 02 Apr 1998 01:22:40 GMT, sl...@nospam.yermom.org (ript)
wrote:

>In article <35248d1c...@news.usol.com>, gd...@rocketmail.com (Gary) wrote:
>>On Wed, 01 Apr 1998 13:59:16 GMT, do...@dgii.com (Doug) wrote:
>>
>>>gd...@rocketmail.com (Gary) wrote:
>>>

>>>>Nope, I don't see any law of physics that would support that statement
>>>>as far as switching states or speed is concerned. My K6-233 runs at a
>>>>3.2 or 3.3 volt core voltage. some people have *REDUCED* that to 3.0V
>>>>or 2.8 V to reduce the power consumption/heat. How does that *lessen*
>>>>its performance ?? It's still changing states 233 million times per
>>>>second. (233 MHz) and processing data as fast as my K6 at 3.3 volts.
>>>>Dropping the voltage on a CPU *may* lessen its *reliabilty* but it
>>>>doesn't lessen its speed/performance.
>>>

>>>Sorry, but LShaping is correct. Transistors switch faster when they
>>>have more voltage to work with. With a higher voltage, the
>>>transistor's "stored charge" drains-off faster (more current).
>>
>>You need some *basic* electronics courses , tranisistors don't "Store
>>Charges" they either switch states when used as digital switching
>>devices (in our uses for them in CPU's Etc.) from one potential to
>>another, cutoff (No output) = 0 , saturation (full on) = 1 ( That's
>>where the binary notation of 0's and 1's come in) or else in analog
>>applications ( amplification) transistors turn on (usually linearlly)
>>to the amount of base current applied , that's where you see the term
>>"Linear Amplifier" Used in Stereos, radios and other analog devices.
>>
>> What you basically described in your above statement is a
>>"Capcacitor" which holds or stores a charge and not a transistor.
>
>
>the junction of a transitor has a bit of capacitance, enough to cause problems
>at realy high speeds, and it is one ot the MAJOR reasons why processors
>disipate more power at higher clock speeds.....

You snipped off the bottom portion of my message where I addressed the
minute capacitance that's inherent in transistors and/or junction
areas. A higher voltage "stored charge" on a transistor does *not*
translate into faster switching speeds as the above poster alluded to.

Gary

[Gary]

On 2 Apr 1998 03:18:57 GMT, "LShaping" <LSha...@worldnet.att.net>
wrote:

>(stuff clipped)
>
>
>> I can't find the article, but someone suggested a Radio Shack Indoor-
>> Outdoor thermometer with the outdoor probe stuck in the CPU heatsink
>> fins. I just went to get one, and they have a Clearance (at least, the
>> store I went to) on cat. 63-1011 ($9.97), which also has an overtemp
>> alarm. Not too shabby.
>
>GREAT ADVICE

Since I posted the (recent) article thanks for the compliment!!

> if the alarm setting can be high enough to allow for operation
>at full capacity. Sounds like the ideal solution for those of you who have
>expensive CPUs sitting on outdated motherboards.

Some of us (myself included) have up to date motherboards but have
avoided those with temp sense features because of too many stories of
the built in sensor not coming into contact with the CPU (measuring
air temperature instead) , inacurate readings , faulty software for
the temp sense function etc. .... Did you buy a MB with a temp sense
feature and now all the otthers are outdated ? ... I'm impressed.

Gary

[Barney Google]

On Thu, 02 Apr 1998 02:48:37 GMT, gd...@rocketmail.com (Gary) wrote:

>On Wed, 01 Apr 1998 17:13:32 -0500, "Paul A. Jacobi"
><jac...@star.enet.dec.nospam.com> wrote:
>
>>Byron Bodo wrote:
>>> Is it feasible to measure power consumption as an alternative
>>> measure of CPUidle's effectiveness, particularly for K6s?
>>>
>>
>>I have a digital watt meter that measures current drawn from the A/C
>>outlet.
>>On a Digital Celebris GL6200 (Pentium Pro 256K), I measured that CPUIDLE
>>drops
>>power consumption approximately 5 watts. This isn't a very big %
>>savings in
>>power,
>
>CPUIdle wasn't designed to save energy, it was designed to reduce
>*heat* given off by lots of current fast processors. In many instances
>(depending on the application) people that formerly had to use a huge
>heatsink and/or fan or keep their case cover off to keep the CPU from
>overheating can now use a normal fan/heatsink and keep the case cover
>on. CPUIdle just stops a number of clock cycles between interrupts,
>basically when there's no data to process (between interrupts) the
>processor is sleeping (very little power drawn) and thus less *heat*

Thus less energy being used dumb shit.

[Gary]

On Thu, 02 Apr 1998 05:04:52 GMT, n...@my.place (Barney Google) wrote:

>On Thu, 02 Apr 1998 02:48:37 GMT, gd...@rocketmail.com (Gary) wrote:
>
>>On Wed, 01 Apr 1998 17:13:32 -0500, "Paul A. Jacobi"
>><jac...@star.enet.dec.nospam.com> wrote:
>>
>>>Byron Bodo wrote:
>>>> Is it feasible to measure power consumption as an alternative
>>>> measure of CPUidle's effectiveness, particularly for K6s?
>>>>
>>>
>>>I have a digital watt meter that measures current drawn from the A/C
>>>outlet.
>>>On a Digital Celebris GL6200 (Pentium Pro 256K), I measured that CPUIDLE
>>>drops
>>>power consumption approximately 5 watts. This isn't a very big %
>>>savings in
>>>power,
>>
>>CPUIdle wasn't designed to save energy, it was designed to reduce
>>*heat* given off by lots of current fast processors. In many instances
>>(depending on the application) people that formerly had to use a huge
>>heatsink and/or fan or keep their case cover off to keep the CPU from
>>overheating can now use a normal fan/heatsink and keep the case cover
>>on. CPUIdle just stops a number of clock cycles between interrupts,
>>basically when there's no data to process (between interrupts) the
>>processor is sleeping (very little power drawn) and thus less *heat*
>
>Thus less energy being used dumb shit.

If you'd read the prior post Horse's Ass you'd see that the poster
said he didn't see much energy savings. My response was that CPUIdle
was designed to REDUCE *HEAT* , it WASN'Tdesigned to neccessarily be
a GREEN (Energy Saving) FEATURE Try Reading and Comprehension 101.
Windows/most BIOS's already *have* green features for energy saving
purposes.

Gary

[Zhenika]

(Regarding the article from Thu, 02 Apr 1998 03:27:20 GMT, by Gary )

I wonder, did anybody try to measure the temperature change
with/w-out Idle when playing Quack II or some other CPU-consuming
games.

You say if the CPU is loaded, it should not make a big change. But
still interesting, what the experiment would show - I mean, just
_how much_ loaded the CPU _realy_ is during gaming?

Regards,
Evgenij

[Rolf Magnus]

Gary wrote:
> >>You need some *basic* electronics courses , tranisistors don't "Store
> >>Charges" they either switch states when used as digital switching
> >>devices (in our uses for them in CPU's Etc.) from one potential to

YOU need some basic electronics curses.

> minute capacitance that's inherent in transistors and/or junction
> areas. A higher voltage "stored charge" on a transistor does *not*
> translate into faster switching speeds as the above poster alluded to.
>

Now what? Once you say a transistor don't store charges and in the next
msg you say they do?
You should think about what you say before you do it.

[Doug]

gd...@rocketmail.com (Gary) wrote:

>
>Nonsense, ( you need to study a bit more electronics)
>

Gee, my Bachelor's degree in electrical engineering isn't enough
study? (I didn't want to say that, but he made me.)

You are wrong. You are the one who needs to study. Do you think that
a transistor is some magic device that doesn't have resistance,
cpacitance, or inductance? There is a "stored charge" in saturated
transitors, that needs to be drained before the transistor can change
states. Higher voltage facilitates this.

LOWER SUPPLY VOLTAGE = SLOWER TRANSISTOR SWITCHING.

Do you know anything about solid-state switching design? Do you know
anything about the different logic families (i.e. TTL, FAST, HCT) and
understand the design trade-offs between them?

Sorry, but when some doorknob tells me that I need to start studying
"basic" electronics, it gets my gander up.

[grri...@ingr.com]

I was very skeptical at first about this program, but I loaded it
under Win95 and did a comparison. WinBench 97 scores changed in some
respects, I can post which if you are curious, but my system seems to
run the same overall. I let the computer sit idle for an hour and the
temp was 41 C. I then shut down and let the temp drop for a few hours
and then loaded the CpuIdle, after 1 hour my system was only 35 C.
After hard 3D gaming, the temp peaks at 42 C, but I have noticed that
if I shut down the game and walk away, the system cools much faster
(the temp drop is much quicker) Pretty good program. I looked for the
difference in voltages on the different lines, and there was no
change. There was also no change in fan speed. What power supplies
have this behavior after running the program?

>>CpuIdle runs a HLT command in an idle priority thread under Win95/98. That
>>allows modern microprocessors to save power and stay cool. Great for
>>overclocking.
>>
>>CpuIdle's is also the best and most complete CPU optimizer for Win95/98
>that
>>exists. Most CPUs include performance increasing options that are disabled
>>by default. CpuIdle activates them all!
>>
>>What's new in Version 2.5 (March, 29th, 1998):
>>
>>- program code completely rewritten
>>- CpuIdle now completely and automatically optimizes your CPU for
>>performance enhancement
>>- removed cpuidle.vxd
>>- removed -loadonly switch since it didn't offer failure protection
>>If you need -loadonly try the loadonly.exe standalone program
>>- added far better power saving for P5 and P5MMX (1.5A..1.8A less)
>>- added -quit switch to exit CpuIdle after CPU optimization
>>- removed all Cyrix specific switches from 2.4 (default is now enabled)
>>- removed -noID switch (nobody used this anyway)
>>- removed -wait switch (no longer needed)
>>- added new switches to selectively disable optimization
>>- workaround for Cyrix and IDT clock rate bug
>>- added CPUID for Cyrix CPUs even if disabled
>>- DOS based optimizers like set6x86 no longer needed for complete Cyrix
>>support
>>- added new direct95.vxd driver to get access to
>>
>>Also included in the release archive is -= DosIdle 2.00 =- by Marton Balog
>>Used in conjunction with CpuIdle, DosIdle and CpuIdle provide a complete
>>power saving and optimization solution both for Win95/98 and DOS, even
>>incompatible DOS programs in Win95 DOS boxes are supported. Please notice
>>that DosIdle is not a CpuIdle addon but a standalone program. These
>programs
>>are best used together and packaged for greater convenience.
>>
>>Andreas Goetz
>>mailto:go...@stud.uni-hannover.de
>>http://www.stud.uni-hannover.de/~goetz
>>A Sysops' Work Is Never Done.
>>
>>
>>
>>
>>
>
>