Re: Mersenne: Overheating!

2003-03-21 Thread nrussell


--On Friday, March 21, 2003 10:13 PM + Elias Daher 
<[EMAIL PROTECTED]> wrote:

I'm back to factoring now, and my cpu temp is never reaching 60°C (over
65 with LL) and the board 50°C (over 60 with LL) which I consider not too
high...
I've had crashes with temps in the high sixties and low seventies before. 
Personally this tends to happen when I turn up the heat in my dorm room and 
close the window when wearing summer PJs.

My highly non-technical solution is not to wear summer PJs, or to wear a 
blanket or robe over them :).

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


Re: Mersenne: Overheating!

2003-03-21 Thread Elias Daher


Hi all again,
I'm back to factoring now, and my cpu temp is never reaching 60°C (over 65 
with LL) and the board 50°C (over 60 with LL) which I consider not too 
high...
Doesn't it mean that the factoring code can be optimized? It's not taking 
advantage of all "parts" of the cpu apparently... Are some "parts" of the 
cpu useless for factoring while they are usefull for LL testing?
Regards,
Epsilon (looking for a new factoring algorithm... wish me good luck!)

_
Tired of spam? Get advanced junk mail protection with MSN 8. 
http://join.msn.com/?page=features/junkmail

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


Re: Mersenne: Overheating!

2003-03-19 Thread Steinar H. Gunderson
On Wed, Mar 19, 2003 at 10:38:52AM -0800, John R Pierce wrote:
> hey, speaking of... anyone care to point me to something I can understand on
> how to configure lm_sensors in Linux to work with an Intel server board?
> I'm getting no readings...

Try "sensors-detect".

/* Steinar */
-- 
Homepage: http://www.sesse.net/
_
Unsubscribe & list info -- http://www.ndatech.com/mersenne/signup.htm
Mersenne Prime FAQ  -- http://www.tasam.com/~lrwiman/FAQ-mers


Re: Mersenne: Overheating!

2003-03-19 Thread John R Pierce
> P4 1.8A, retail box HSF, 50C
> P4 2.53B, Zalman CuAl HSF, <2000 rpm, 44C
> P4 2.66B, Zalman CuAl HSF, <2000 rpm, 45C

and to add a data point...

  P4 2.53B, Intel retail HSF, 46C (115F)

this in an Antec plus1080 case, with 2 case fans running at a very slow
speed, the 2 motherboard temp sensors are reading 38C (100F).  If I hook the
case fans up to straight 12V rather than the speed reduced fan power coming
out of the Antec power supply, the case temp drops 5C or more in a matter of
seconds, and the CPU temp tracks.

hey, speaking of... anyone care to point me to something I can understand on
how to configure lm_sensors in Linux to work with an Intel server board?
I'm getting no readings...

# sensors
Can't access /proc file
/proc/sys/dev/sensors/chips or /proc/bus/i2c unreadable;
Make sure you have done 'modprobe i2c-proc'!
# modprobe i2c-proc
# sensors
No sensors found!


(this is with Redhat 8 on a E7501 based dual xeon server board, the
SE7501WV2, kernel linux-2.4.18-27.8.0)


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


Re: Mersenne: Overheating!

2003-03-19 Thread Brian J. Beesley
On Wednesday 19 March 2003 00:10, Elias Daher wrote:
> Thanks all, actually I was thinking of opening the case, (but too lazy to
> do it...) and the weird thing now is that the temperature is stable with
> Prime95 running with other applications, and it's not reaching the level it
> was reaching two days ago and I'm still testing the same number... It's

Odd. I wonder if a fan stalled. Possibly even the case fan, if you have only 
one fitted.

> still hot though, the CPU is running at 65°C and the board at 55°C...
> Anyway, the P4 is tough, it can handle it! (Once the CPU temp reached 85°C
> for more than 10 minutes cause it was running without a fan!!!)

AFAIK the Intel spec for thermal throttling is 75C.

However the P4's I'm running are all reasonably cool:

P4 1.8A, retail box HSF, 50C
P4 2.53B, Zalman CuAl HSF, <2000 rpm, 44C
P4 2.66B, Zalman CuAl HSF, <2000 rpm, 45C

Zalman HSF comes with a rheostat which allows you to vary the fan speed from 
(approx) 1500 - 3000 rpm - below 2000 rpm it's highly unlikely that you will 
be able to hear the fan, so this HSF is quiet as well as effective. The only 
change I made to Zalman's installation instructions was to use Arctic Silver 
thermal compund instead of the small tube of gloop included in the Zalman 
kit. Note, I'm not saying Zalman's gloop is useless, but Arctic Silver is 
usually a couple of degrees cooler than other stuff.

A decently ventilated case helps, too ...

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


Re: Mersenne: Overheating!

2003-03-18 Thread Imad R. Faiad
-BEGIN PGP SIGNED MESSAGE-
Hash: RIPEMD160

Hello again Elias,

65°C for a williamette P4 is a normal temperature when run at full load
(i.e. Prime95), with a stock heat sink.

For comparison my 2.4GHz Northwood CPU, overclocked to 2.7GHz, with stock
heatsink/fan, the temperature reads as 57.5C.

Happy number crunching,

Best Regards

Imad R. Faiad


On Wed, 19 Mar 2003 00:10:50 +, you wrote:
ac
>
>Thanks all, actually I was thinking of opening the case, (but too lazy to
>do  it...) and the weird thing now is that the temperature is stable with 
>Prime95 running with other applications, and it's not reaching the level
>it  was reaching two days ago and I'm still testing the same number...
>It's 
>still hot though, the CPU is running at 65°C and the board at 55°C... 
>Anyway, the P4 is tough, it can handle it! (Once the CPU temp reached 85°C
> for more than 10 minutes cause it was running without a fan!!!)
>Regards...

-BEGIN PGP SIGNATURE-
Version: 6.5.8ckt http://www.ipgpp.com/

iQEVAwUBPne2gbzDFxiDPxutAQP99Qf8CLLr4Pj9Cf3rHU6qvcWe/3KUqI6jXYjh
FjG1akYO9jrSD1xqtCLaFvGHnt77YcB0uGZhpVJaArpYv0Ws6bK9hckWmHNrUAaq
bPK3N+GBomXIUJ5PugrUqTpuBijFKV0U/i/rl9H5eDDRdDWOBNSjWPW8Sfdv0uIK
g/fktAajk8k1gqqPNwe11PHRCkZVlhIDxjT/6yeUfP1Qm0tbDrxVCL+LdEKs5iVX
R0ps2ZwLe4US/H/p4vxMpfuGIGF8wGPLFhZ98k2m7/h9tN1seO3QBkyNxr1iuT5p
Vh6PHylSQu9rpUxzdHS5FyHpH74c1is7R84Nex6J7Ky2mKHpCPTJ9g==
=+ke1
-END PGP SIGNATURE-

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


Re: Mersenne: Overheating!

2003-03-18 Thread Elias Daher
Thanks all, actually I was thinking of opening the case, (but too lazy to do 
it...) and the weird thing now is that the temperature is stable with 
Prime95 running with other applications, and it's not reaching the level it 
was reaching two days ago and I'm still testing the same number... It's 
still hot though, the CPU is running at 65°C and the board at 55°C... 
Anyway, the P4 is tough, it can handle it! (Once the CPU temp reached 85°C 
for more than 10 minutes cause it was running without a fan!!!)
Regards...

_
The new MSN 8: advanced junk mail protection and 2 months FREE* 
http://join.msn.com/?page=features/junkmail

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


Re: Mersenne: Overheating!

2003-03-18 Thread Imad R. Faiad
-BEGIN PGP SIGNED MESSAGE-
Hash: RIPEMD160

Hello Elias,

I guess you must have one of those williamette P4's, these tend to warm up
a bit.  Since you had no such problems for the past three months, I infer,
I would recommend that you open he case and dust off the heat sink and
chipset. Dust tends to accumulate over time. Should that not help, I would
look into the power supply.

Hope the above helps,

Best Regards,

Imad R. Faiad

On Mon, 17 Mar 2003 21:20:33 +, you wrote:

>
>Hi
>I had let my computer work on factoring for a while, and I switched back
>to  LL testing two days ago, and last night my computer was beeping all
>the time  because the cpu and the board were overheating... There should
>be no problem  with my hardware, I got good fans that I bought 3 months
>ago, and everything  was fine until yesterday... I have a P4 1.5 Ghz...
>There's something weird I  noticed but I don't know if it's a factor and
>it is that SiSoft Sandra 
>indicates the voltage sensor is detecting -8.7V and -3V for the -12 V and
>-5  V respectively...
>Does anyone have an idea? (sure! ;-) ) Does version 23.2.1 of prime95
>(which  I'm using) use the cpu more efficiently (and overheating it as a 
>consequence) than the previous versions which weren't causing any problems
> on my current configuration?

-BEGIN PGP SIGNATURE-
Version: 6.5.8ckt http://www.ipgpp.com/

iQEVAwUBPneR1LzDFxiDPxutAQPuwwf+Lbky2uIogqAhh5OVlYtbJSiT4lWHVXnv
5AIJWw2K9VIZ1YvTeDyggXIVE115+gd0IvjNQwaviIYsEk0Zfl62el+svrbeKtKJ
5htJ4ZnJ7r8tJKgM3P0kvrAyTglgQhm4QVxlYg+as4AUkWB/iW3b2ijmlhZn8oWv
VQlRXxpLVU7+tP8ju9EOPV9H/4sHmpH5hoMlT1AQiEjIDY5tFmv5AuG8HZzFIIaX
Jnq7HNMSVg/XNnsbu/+tRprrNnrH91qVYtN9JddKzSoK5DmwYJP+8KLgoy+fXdy2
tV/ztFjk4sePdoo2EBLJPBqD6lG723UdSQW3nw3KhLyDwQAGzrJflw==
=CJL9
-END PGP SIGNATURE-

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


Re: Mersenne: Overheating!

2003-03-18 Thread Brian J. Beesley
On Monday 17 March 2003 21:20, you wrote:
> Hi
> I had let my computer work on factoring for a while, and I switched back to
> LL testing two days ago, and last night my computer was beeping all the
> time because the cpu and the board were overheating... There should be no
> problem with my hardware, I got good fans that I bought 3 months ago, and
> everything was fine until yesterday... I have a P4 1.5 Ghz... There's
> something weird I noticed but I don't know if it's a factor and it is that
> SiSoft Sandra indicates the voltage sensor is detecting -8.7V and -3V for
> the -12 V and -5 V respectively...

Weird. Did you try a different version of Sandra? I suspect the version 
you're using doesn't understand the chipset properly.

> Does anyone have an idea? (sure! ;-) ) Does version 23.2.1 of prime95
> (which I'm using) use the cpu more efficiently (and overheating it as a
> consequence) than the previous versions which weren't causing any problems
> on my current configuration?
>
I doubt Prime95 version has much to do with it. The P4's I have are all 
Northwood processors (1.8A, 2.53 & 2.66); the slower two have been running 
since v21 & I haven't seen any significant jumps in CPU temp with version 
(note, I use mprime rather than Prime95).

It's certainly true that running LL tests (or DC assignments) will drive the 
chip hotter than running factoring, because the parts of the chip involved in 
SSE2 will be fully active instead of more or less idle.

Other things which could be a significant factor:

(a) chipset setup - usually you can mess around with fan speeds and/or 
thermal throttling to change the balance between speed/power consumption and 
noise/battery life - particularly battery life of notebooks. Note, sometimes 
this setup can be affected by power saving parameters settable from within 
the running operating system as well as through the BIOS at system boot time.

(b) systems which have been running for a while often accumulate dust, this 
can get stuck in between heatsink fins & reduce efficiency of cooling.

However, fans do sometimes fail (even under 3 months age); my guess is that 
your CPU cooler fan has deceased. The thermal control circuitry embedded in 
the P4 does at least keep the system limping along with a failed CPU cooler 
fan. If the heatsink itself is halfway decent, and the case is properly 
ventilated, overheating could well fail to occur unless something heavy like 
LL testing is running.

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


Mersenne: Overheating!

2003-03-17 Thread Elias Daher
Hi
I had let my computer work on factoring for a while, and I switched back to 
LL testing two days ago, and last night my computer was beeping all the time 
because the cpu and the board were overheating... There should be no problem 
with my hardware, I got good fans that I bought 3 months ago, and everything 
was fine until yesterday... I have a P4 1.5 Ghz... There's something weird I 
noticed but I don't know if it's a factor and it is that SiSoft Sandra 
indicates the voltage sensor is detecting -8.7V and -3V for the -12 V and -5 
V respectively...
Does anyone have an idea? (sure! ;-) ) Does version 23.2.1 of prime95 (which 
I'm using) use the cpu more efficiently (and overheating it as a 
consequence) than the previous versions which weren't causing any problems 
on my current configuration?

_
Add photos to your messages with MSN 8. Get 2 months FREE*. 
http://join.msn.com/?page=features/featuredemail

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