cputemp over 100c, 4*sp2004 benchable on Air? QX6700es B3 say yes !

Printable View

10-15-2006, 06:06 AM
VictorWang

1 Attachment(s)

cputemp over 100c, 4*sp2004 benchable on Air? QX6700es B3 say yes !

random pick a QX6700es B3 and asked my partner send to my Hotel,
tested on a un-moded P5W64 and that Bad bad Zalmen 9500LED.
ambient=25c :woot:

astonished me so much, asus probe read=95c, so coretemp must be over 100c+ :eek:
but...still benchable 4*sp2004 :woot: :clap:
volt. can be lower, but I haven't more time for testing, busy these days.
just let vcore=defult :D
10-15-2006, 06:07 AM
VictorWang

and, 1.43v 3.75G pi_32m ok, seems good.

http://www.xtremesystems.org/forums/...1&d=1160920718
10-15-2006, 06:11 AM
Dumo

Nice cpu Victor:toast:
Thats inferno:)
10-15-2006, 06:17 AM
Devious

I just returned my P5W64 because it was reporting stuipidly high temps.
10-15-2006, 06:24 AM
VictorWang

Quote:

Originally Posted by Devious

I just returned my P5W64 because it was reporting stuipidly high temps.

no, I think it's the right temp.
only about 10c lower than CoreTemp shows.

if use a Giga's mobo, shows about 15c~20c lower than coretemp!
10-15-2006, 06:45 AM
HamidFULL

Nice Victor! you need watercooling
10-15-2006, 07:24 AM
lawrywild

Quote:

Originally Posted by Devious

I just returned my P5W64 because it was reporting stuipidly high temps.

lol? you returned a mobo because you didn't like what temps it reported?
AHAHAHA :slap:
10-15-2006, 07:33 AM
afireinside

Quote:

Originally Posted by VictorWang

no, I think it's the right temp.
only about 10c lower than CoreTemp shows.

if use a Giga's mobo, shows about 15c~20c lower than coretemp!

No way it's loading 94c. Doesn't the solder on the IHS melt in the 70s?

Also I thought dual orthos still only loaded 2 cores?
10-15-2006, 07:44 AM
Devious

Quote:

Originally Posted by lawrywild

lol? you returned a mobo because you didn't like what temps it reported?
AHAHAHA :slap:

It was showing over 95deg when priming for a few seconds @ 3.6Ghz with 1.375v ! and probably causing cpu throttling. I wanted to push 1.6v through the chip and get well over 4Ghz, could you imagine the temps then ? Complete BS, faulty in my opinion so back it went.
10-15-2006, 07:50 AM
VictorWang

Quote:

Originally Posted by afireinside

No way it's loading 94c. Doesn't the solder on the IHS melt in the 70s?

Also I thought dual orthos still only loaded 2 cores?

@.@

HAVE YOU SAW TASKMANAGER SHOWS 4THREADS 100% LOADED ? :cool:
SOLDER ON IHS MELT IN 70S ? WHERE HAVE YOU GET SUCH INFO. ? :slobber:
10-15-2006, 07:52 AM
ScythedBlade

Nope ... solder doesn't melt at 70s .... it melts at much higher ... i mean COME ON ... WATER BOILS at 100C .... solder so does not melt in 70 C
10-15-2006, 08:25 AM
Millyons

Quote:

Originally Posted by Devious

It was showing over 95deg when priming for a few seconds @ 3.6Ghz with 1.375v ! and probably causing cpu throttling. I wanted to push 1.6v through the chip and get well over 4Ghz, could you imagine the temps then ? Complete BS, faulty in my opinion so back it went.

did u try any other mobos, i also didn't believe my temps of my 6600 on the p5b dlx. I have a 30c delta, with SI-120 and 1.425 3.4ghz i got idle of 42 ish and load of over 70 but i also tried it on the Gigabyte s3 and the temps were very simular, ofcourse measured with core temp, it could be that u just have a "Hot running" cpu, either bad IHS or not making a good contact between the core and the IHS, or mabey (i dont know how plausabe taht is) the internal cpu probes aren't calibrated, it would be really cool if the reason for high temp was figured out
10-15-2006, 08:28 AM
Mykou

Quote:

Originally Posted by VictorWang

no, I think it's the right temp.
only about 10c lower than CoreTemp shows.

if use a Giga's mobo, shows about 15c~20c lower than coretemp!

what about TAT temps ?

core temp more accurate than TAT ?

core temp shows me 3° higher than orthos
10-15-2006, 08:49 AM
Dani

My P5W64WS Pro shows this: ;)

- IDLE
- LOAD

as you can see, Probe shows ~3'C more than CoreTemp..
10-15-2006, 09:40 AM
Devious

Quote:

Originally Posted by Millyons

did u try any other mobos, i also didn't believe my temps of my 6600 on the p5b dlx. I have a 30c delta, with SI-120 and 1.425 3.4ghz i got idle of 42 ish and load of over 70 but i also tried it on the Gigabyte s3 and the temps were very simular, ofcourse measured with core temp, it could be that u just have a "Hot running" cpu, either bad IHS or not making a good contact between the core and the IHS, or mabey (i dont know how plausabe taht is) the internal cpu probes aren't calibrated, it would be really cool if the reason for high temp was figured out

3 other boards all with normal temps, I currently have a P5B Dlx, previously I have had a bad axe and a P5W DH, see my sig, on my P5W DH duel priming at 3.94GHz temps were 60deg in core temp.

The P5W64 was bad and others have reported the same.
10-15-2006, 10:13 AM
Lestat

victor at 1.32v even with a stock cooler whch is worse than that horrible zalman temps still wouldnt be almost 100c. no they wouldnt its not possible.

the only, and ONLY explanation is that the heatsink is making 10% contact and the Zalman cooling system of the heatsink is just totally fuxxored.

the thermal protection of the cpu would have kicked in anyways and turned your system off.

you can turn that off in the bios but the cpu itself still has a thermal limit and it will shut down. and you have bypassed that thermal limit already.

you might as well throw that cpu away its toast now.
10-15-2006, 11:04 AM
Millyons

Quote:

Originally Posted by Devious

3 other boards all with normal temps, I currently have a P5B Dlx, previously I have had a bad axe and a P5W DH, see my sig, on my P5W DH duel priming at 3.94GHz temps were 60deg in core temp.

The P5W64 was bad and others have reported the same.

man what was core temp or tat showing, were they also alot higher than on the other boards, cause if they were that just showe that they also rely on mobo readouts not just cpu, as some ppl are are claiming
10-15-2006, 11:10 AM
Millyons

Quote:

Originally Posted by Lestat

.....
you can turn that off in the bios but the cpu itself still has a thermal limit and it will shut down. and you have bypassed that thermal limit already.

you might as well throw that cpu away its toast now.

man how do you figure that,

i purposly ran my conroe to see when it was gonna shut off and it was 111c measured by probe, tat or coretemp dont show above 85, and ofcourse the cpu is fine and running 24/7

oh ya i also continously ran it above 85c (tat, core temp) with orthos for few hours and its fine
10-15-2006, 11:15 AM
Kjaks

Quote:

Originally Posted by Millyons

man how do you figure that,

i purposly ran my conroe to see when it was gonna shut off and it was 111c measured by probe, tat or coretemp dont show above 85, and ofcourse the cpu is fine and running 24/7

oh ya i also continously ran it above 85c (tat, core temp) with orthos for few hours and its fine

The protection is triggered from the digital measurer in the cpu, not from what the motherboard/apps register. If the Kentsfield was throttling sp 2004 would report errors and stop.
10-15-2006, 11:16 AM
ochw

thats hot man:)
10-15-2006, 11:54 AM
Devious

Quote:

Originally Posted by Millyons

man what was core temp or tat showing, were they also alot higher than on the other boards, cause if they were that just showe that they also rely on mobo readouts not just cpu, as some ppl are are claiming

Like I said, core temp hit 95 deg in seconds on the P5W64, on the other boards it didnt go higher than around 60 deg with much higher vcore and clocks.

It would also show around 60 deg in hardware monitor in bios where the other boards wouls show 25-30deg.
10-15-2006, 12:22 PM
sdumper

Quote:

Originally Posted by Devious

3 other boards all with normal temps, I currently have a P5B Dlx, previously I have had a bad axe and a P5W DH, see my sig, on my P5W DH duel priming at 3.94GHz temps were 60deg in core temp.

The P5W64 was bad and others have reported the same.

Im also showing pretty high temps with this mobo on phase I thought my phase needed servicing since I was showing -20c idle and I am normally way lower than that with my Chilly ... hmm...maybe ill do some more experimenting using different cooling and chips. I appreciate you bringing this up...
10-15-2006, 12:28 PM
sdumper

Quote:

Originally Posted by Millyons

man how do you figure that,

i purposly ran my conroe to see when it was gonna shut off and it was 111c measured by probe, tat or coretemp dont show above 85, and ofcourse the cpu is fine and running 24/7

oh ya i also continously ran it above 85c (tat, core temp) with orthos for few hours and its fine

I second that...

I have had a couple of occasions where my phase went awry and temps hit a peak and the CPU shut down (i always bypass temps in bios). Well my e6700 suffered no ill effects and still benches fine in the 4.3ghz range and folds 24x7 at 4ghz...

So toast???...I think not :D
10-15-2006, 12:39 PM
largon

94ºC?
No way.

Quote:

Originally Posted by VictorWang

SOLDER ON IHS MELT IN 70S ? WHERE HAVE YOU GET SUCH INFO. ? :slobber:

Quote:

Originally Posted by ScythedBlade

Nope ... solder doesn't melt at 70s .... it melts at much higher ... i mean COME ON ... WATER BOILS at 100C .... solder so does not melt in 70 C

Intel IHS-to-die solder does actually melt at about 80ºC.

The solder I'm talking about ofcourse is not tin solder but a alloy composing mostly of gallium with some indium and tin to make it solid under ~80ºC.

How do I know?
Well, I've melted it...
10-15-2006, 01:38 PM
Millyons

Quote:

Originally Posted by Kjaks

The protection is triggered from the digital measurer in the cpu, not from what the motherboard/apps register. If the Kentsfield was throttling sp 2004 would report errors and stop.

no it wouldn't, it would just keep running, you can see it throttle with cpuz, or rightmark, throtle watch doesnt show it, i know i have tried it
10-15-2006, 01:42 PM
fhpchris

Quote:

Originally Posted by Millyons

man how do you figure that,

i purposly ran my conroe to see when it was gonna shut off and it was 111c measured by probe, tat or coretemp dont show above 85, and ofcourse the cpu is fine and running 24/7

oh ya i also continously ran it above 85c (tat, core temp) with orthos for few hours and its fine

Same here, 112c on Asus probe on P5WDG2 WS PRO shuts off :)
10-15-2006, 01:44 PM
sierra_bound

Kentsfield has a higher Max Tcase temp than Conroe. 100C. Whether 94C is accurate is not the issue. Kentsfield does get extremely hot under full load. I was getting temps in the 80's running quad-SP2004 with VCore at just 1.43v.
10-15-2006, 05:23 PM
irev210

Quote:

Originally Posted by sierra_bound

Kentsfield has a higher Max Tcase temp than Conroe. 100C. Whether 94C is accurate is not the issue. Kentsfield does get extremely hot under full load. I was getting temps in the 80's running quad-SP2004 with VCore at just 1.43v.

yah no kidding... we are going to see 100+ dollar air coolers become the norm here.

At stock vcore, kentfield is fine, but once you start loading it up with vcore, things get out of hand.

I saw ~300 watt power draw on a Kentfiend overclock to 4.45ghz with a P5W DH Dlx, an ati x300 video card and 2gb of D9GMH w/ 2.45vdimm

This was at idle!

With stock vcore, it pulls closer to 100 with the same stuff at idle.
10-15-2006, 05:32 PM
sierra_bound

One thing people need to be aware of is that some boards have a bios bug when using Kentsfield. When you change the multiplier in the bios, only one of the dies sometimes recognizes the change. This is one reason why some boards have problems with high FSB. One die thinks it's running at x10, while the other thinks it's at x9. The P5B Dlx is one of the few boards that changes the multipliers correctly on both dies. VictorWang was the first to point out this bug.

http://www.xtremesystems.org/forums/...d.php?t=115425

The problem is not just limited to the board he was using.
10-15-2006, 05:38 PM
ooeric

burn baby burn!
i wonder if i put that cpu into my WC loop.. will i see steam bubbles coming up to the fillport line? tehe
melt that price of couple hundred $ silicone. haha
but boy. those new quads are clocking better then duos.
hasnt anyone noticed that?
10-15-2006, 05:45 PM
sierra_bound

Kentsfield clocks better than some Conroes, but not all. The problem when you start raising VCore is the heat generate will begin to slow down programs. Most people think when a CPU gets too hot, it will simply reboot or the motherboard will shut down. With Kentsfield, you can still run Windows with high VCore, but apps will be sluggish.
10-15-2006, 08:07 PM
Dumo

Probly these 2X850Watter will comfortably suffice for that beast of cpu + CF/SLI oc'd...

Just trying to cover all bases before switch to Quad...

http://img183.imageshack.us/img183/5...shot084fj8.jpg
10-16-2006, 12:48 AM
Durzel

I've got you all beat, I've seen 255 degrees on my E6700. ;)
10-16-2006, 02:23 AM
death metal

That's one high temp man!!!
10-16-2006, 07:37 AM
VictorWang

haha, I think most of u guys are mis-understood my meaning.
my point is ... same as Sierra and dumo, kentsfield can running/bare at a very high temp. asus probe shows the right Core Temp(p5w64/p5b-dl/ws pro).
btw, zalman 9500led is not good as u guys thought. in fact, it's really a bad cooler with bad mounting design.
I've compared with Tower120 in same condition, about 15~20c loaded temp lower than 9500LED. :D
all we need for kentsfield, is a 50usd cooler and a 500w+ PSU.
QX6700+X1950XTX+RaptorX+2GB power consumption=200w when running quad-sp2004/3dmark with auto vocre/vgpu. :D:D
10-16-2006, 07:59 AM
eva2000

Vic maybe try a Scythe Infinity + high cfm fan

On D975XBX Badaxe r304

955XE @16x266fsb = 4266Mhz stock speeds at 1.5v bios vcore (1.441v bios)

Scythe Infinity with 120x25mm Thermaltake 81cfm 0.40AMP fan

Bios value => Intel Desktop Utilities value (windows)
CPU idle = 74C => 51C
System Zone 1 idle = 39C => 34C
System Zone 2 idle = 31C => 31C

Intel Desktop Utilities value (windows)
CPU load = 79-80C
System Zone 1 load = 41-42C
System Zone 2 load = 32-33C

Room temp = 25.4C

Idle
http://fileshosts.com/intel/Intel975...pu_idle_tn.png

Load (click image for larger image)
http://fileshosts.com/intel/Intel975...e_load_tnn.jpg

955XE @4266Mhz at 1.441v = ~220 watts

Vic, how does QX6700 B3 compare to 955XE in terms of heat/watts ?
10-16-2006, 08:05 AM
kiwi

It is hotter but are do you really expect Intel will include Tower/BT/Infinity as a boxed cooler? As long as your heatpipes are OK to touch it should be just fine :)

All times are GMT -8. The time now is 08:45 AM.

XtremeSystems