Average CPU Temperatures

alexs3d2

Distinguished
Feb 4, 2006
23
0
18,510
I was recently interested to see the Temperature levels of my old, but still very capable, Northwood 2.8GHz and I was surprised to see the low average Temperature...

I mean, hearing the critisism against intel about the Pentium 4 these last two years, i was left with the impression these chips should really be hot.

Anyway, just for comparison purposes, I get around 26°C average and under heavy load a maximum of 38°C. I am using a stock intel cooler running with ASUS Q-FAN set at 11/16 (pretty quiet at these speeds, unlike Athlon stock coolers I've heard) aka. 1962rpm.

The chipset is always running hotter, around 6°C more average and 2°C under load. But this I expected, because it only has a heatsink, no fan.

Btw, I am running on an ASUS P4P800-SE.

So what are your numbers?
 

1Tanker

Splendid
Apr 28, 2006
4,645
1
22,780
I was recently interested to see the Temperature levels of my old, but still very capable, Northwood 2.8GHz and I was surprised to see the low average Temperature...

I mean, hearing the critisism against intel about the Pentium 4 these last two years, i was left with the impression these chips should really be hot.

Anyway, just for comparison purposes, I get around 26°C average and under heavy load a maximum of 38°C. I am using a stock intel cooler running with ASUS Q-FAN set at 11/16 (pretty quiet at these speeds, unlike Athlon stock coolers I've heard) aka. 1962rpm.

The chipset is always running hotter, around 6°C more average and 2°C under load. But this I expected, because it only has a heatsink, no fan.

Btw, I am running on an ASUS P4P800-SE.

So what are your numbers?
Heat wasn't an issue until Prescott came along. I had similar temps on my 3.0C, and i have a 2.4C@2916 currently and it also jives with your numbers.
 

306maxi

Distinguished
Feb 7, 2006
679
0
18,980
I was recently interested to see the Temperature levels of my old, but still very capable, Northwood 2.8GHz and I was surprised to see the low average Temperature...

I mean, hearing the critisism against intel about the Pentium 4 these last two years, i was left with the impression these chips should really be hot.

Anyway, just for comparison purposes, I get around 26°C average and under heavy load a maximum of 38°C. I am using a stock intel cooler running with ASUS Q-FAN set at 11/16 (pretty quiet at these speeds, unlike Athlon stock coolers I've heard) aka. 1962rpm.

The chipset is always running hotter, around 6°C more average and 2°C under load. But this I expected, because it only has a heatsink, no fan.

Btw, I am running on an ASUS P4P800-SE.

So what are your numbers?

These are the temps from my old P4 2.8 witha Coolermaster Hyper6 on it

Chilly.JPG


Your temps are nothing to worry about.
 

m25

Distinguished
May 23, 2006
2,363
0
19,780
I was recently interested to see the Temperature levels of my old, but still very capable, Northwood 2.8GHz and I was surprised to see the low average Temperature...

I mean, hearing the critisism against intel about the Pentium 4 these last two years, i was left with the impression these chips should really be hot.

Anyway, just for comparison purposes, I get around 26°C average and under heavy load a maximum of 38°C. I am using a stock intel cooler running with ASUS Q-FAN set at 11/16 (pretty quiet at these speeds, unlike Athlon stock coolers I've heard) aka. 1962rpm.

The chipset is always running hotter, around 6°C more average and 2°C under load. But this I expected, because it only has a heatsink, no fan.

Btw, I am running on an ASUS P4P800-SE.

So what are your numbers?

These are very good temps but it may also be a faulty sensor (especially if you don't have another fan blowing out of your case); confirm them by touching the heatsink. My celeron of 2GHz (Northwood) used to run around 35-36°C idle.
 

Grimmy

Splendid
Feb 20, 2006
4,431
0
22,780
I'm on an older Asus board, P4PE, 2.66-533, OC to 2.940 atm (having problems now @ 2.960, Prime95 not stable). I think the vcore on mine gets high, if its reading is correct. 1.6-1.78 Jumps allot. I'm still thinking of dropping a P4 3.0c later on.

My Northy seems to run abit hotter then any of these above posted readings:

Room temp 80-82F
Average Idle is 44C
Load 55C.

Lower Room temps 70F
Idle is 38-39C
load 47-48C

I am going to mod my 7700CU fan, since I'm a lil bored. :lol:

The Zalmans 120m fan's CFM is around 30. I'll be waiting on my FM121(CFM110) fan, should be here tommorow.

Might post up the mod later on, if its worth it.
 

alexs3d2

Distinguished
Feb 4, 2006
23
0
18,510
These are very good temps but it may also be a faulty sensor (especially if you don't have another fan blowing out of your case); confirm them by touching the heatsink. My celeron of 2GHz (Northwood) used to run around 35-36°C idle.

Well, I did add two fans in my case last week, but it has hardly changed the readings. I have a Chieftec Dragon Tower, which is quite big and has numerous openings for fabs on the rear. So even without the extra fans, it never felt hotter inside the case than the room temperature.

If I disable the CPU fan, the temperatures may rise upto 70°C and then the system automatically shuts down. If 70 is the maximum allowed, then the readings should be correct.
 

m25

Distinguished
May 23, 2006
2,363
0
19,780
DON'T DO THAT. Disabling the fan is too much of temp. difference. I said just to touch the heatsink: Shut the PC down with covers open and imediately after that touch the heatsink fins: If you don't notice that they're any hotter than your bodu the reading is OK, If they're body temp (you barely feel the heat) it's around 37-38°C. If like somebody with fever it's around 40, got the concept :wink:
 

Grimmy

Splendid
Feb 20, 2006
4,431
0
22,780
Using the word "if/may" sound like guessing.

I did read an article comparing Northwood with Prescotts throttling feature by shutting off the fan, and just running full load on the HS.

On that test, Northwoods end up getting around 92-98C, if I remember correctly, which tripped the thermal switch, shutting the PC off.

But just keep in mind that all MB's will not accurately report temps. I'd say close to it, but not exact.

Edit:

Though it was based on 3.2 Ghz cores:

Testing Thermal Throttling
 

m25

Distinguished
May 23, 2006
2,363
0
19,780
I'm on an older Asus board, P4PE, 2.66-533, OC to 2.940 atm (having problems now @ 2.960, Prime95 not stable). I think the vcore on mine gets high, if its reading is correct. 1.6-1.78 Jumps allot. I'm still thinking of dropping a P4 3.0c later on.

It looks like some Asus boards are really crap; mine runs all S939 single cores @ 1.50V just to include the older Manchester cores of 1.50V. So my venice 3000+ runs with 0.10-0.15V more than it should, this translates in 3-4°C added to load temps (I get the lower temps with vcore controled by CrystalCPUID).
 
Desktop
Athlon 64 X2 4200+
HSF Speed: 2550 rpm (70%)
Ambient temp 20 C
Idle (1.0 GHz): 25 C
Idle (2.2 GHz): 29-32 C
Full load one core: 39-40 C
Full load both cores: 42-44 C
Maximum observed temp: 46 C

Laptop
Mobile Pentium 4-M 2.20 GHz
HSF trips: 55 C = off, 65 C = fan on 75 C = fan on higher
Ambient temp 20 C
Idle (1.2 GHz): reaches 65 C from cold in ~30 min, fan cycles on once for a minute every 5-6 hours
Idle (2.2 GHz): fan cycles on every ~30 min for a minute
Full load: 66-72 C, fan on constantly
Maximum observed temp: 80 C

That's what my chips run.

EDIT: The voltage on the X2 is 1.10V @ 1.0 GHz, 1.35V @ 2.20 GHz, and on the P4-M it is 1.20 V @ 1.2 GHz, 1.30 V @ 2.20 GHz. Both chips run stock.
 

Grimmy

Splendid
Feb 20, 2006
4,431
0
22,780
I'm on an older Asus board, P4PE, 2.66-533, OC to 2.940 atm (having problems now @ 2.960, Prime95 not stable). I think the vcore on mine gets high, if its reading is correct. 1.6-1.78 Jumps allot. I'm still thinking of dropping a P4 3.0c later on.

It looks like some Asus boards are really crap; mine runs all S939 single cores @ 1.50V just to include the older Manchester cores of 1.50V. So my venice 3000+ runs with 0.10-0.15V more than it should, this translates in 3-4°C added to load temps (I get the lower temps with vcore controled by CrystalCPUID).

Umm.. my P4 requires stock 1.525v. To run it stable OC, I need to raise it to 1.675v for 2.960. Anyhoo.. found out why it was not stable.

Seems I forgot I was trying for 3.1Ghz. The sound card wasn't coming up at that point. So I remember I changed the AGP/PCI frequency to 66mhz. That fixed the sound card problem, but it would freeze up on 3.1Ghz OC.

After all that time I forgot to switch it back to auto. Now it runs stable at 2.960Ghz again. Actually now I run it 1.650v @ 2.940Ghz. Thought screw the extra 20mhz and .025 volts.
 

gw

Distinguished
Apr 15, 2004
38
0
18,530
I have a pentium d 940 on a 975 xbx mb. I have a heating problem too, idle temp is 64 deg c and load can get to 80 deg c. I called Intel, not much help, I did all the checks and upgrades,etc..

I was told to check the amount of thermal grease I used. I did and found out I had way too much. I was told to put only a spot the size of two grains of rice. I did and it now runs at 50 deg c.. I think that is still too hot, but it sure seems to run better.

I may try a new heatsink/fan, using the old Pentium 4 fan now.

So be sure not to over do the thermal grease..gw
 

ZOldDude

Distinguished
Apr 22, 2006
1,251
1
19,280
26C but you did not say what the room temp was.

I used to run my Optron 146 (OC to 3Ghz) on air with a Thermalright SI-120 in a Coolermaster case with 2C over room temps (idle) and 4-5C while under benchmarking/burn in testing for about an hour.

About a week ago I installed another 300GB drive and all the temps went up 1C.

I know the rest of you are talking about Intel but thats not the point here as others with my CPU and OC have higher temps.

The point is that with the correct airflow you can run close to room temps and that does not mean a handfull of fans.

I set up my case to bring cool air in,over the cooler and out the back all in the same path so as to not have warm air swirling about like a tornado inside of it....just 2 fans.
The PSU (PC P&C 510) is not being used as a case cooler so the air out of it is about 1C over room temps. I used posterboard and duct tape to feed it air from the blowhole on top of the case (turned the small fan on top around).
 

m25

Distinguished
May 23, 2006
2,363
0
19,780
Umm.. my P4 requires stock 1.525v. To run it stable OC, I need to raise it to 1.675v for 2.960. Anyhoo.. found out why it was not stable.

I don't have that problem with mine: from 1.80GHz to 2.05GHz it runs rock stable even undervolted (with 1.30V instead of the minimal 1.35). The Venice is a great core, almost brought a new era in OC-ing, where you can have 20-30%+ on stock. You do it and are 100% sure it won't do a scratch to your CPU.
 

myou066

Distinguished
Aug 17, 2006
5
0
18,510
hey this is my first post....

anyway i got myself a northwood 2.66 533fsb as well and i got it up to 3.2ghz sweet. only needed a 5% voltage increase and rean prime 95 for 8 hrs no probs. if u want to oc this cpu further might try goin into the bio and lowering the timings of ur ram coz that mite be the limiting factor. But i hit a wall when trying to do 3.3ghz with resets and lockups..

Oh im also using a thermalright xp 90 on a gigabyte GA-8S655FX-L as well
temps are: 26c idle
53c full load max
 

Dmcbeing

Distinguished
Jun 30, 2006
3
0
18,510
Well i have to be the king of (reported) temps :)
i have a p4 nortwood @ 2.0 ghz.Some time ago i was O/C him @ 2.6 but one day i noticed something !
cpu tmep 110c at the bios
i was like WTF!!!!
I gues i must have burned the thermistor some how :p
No o/c for him anymore.He runs at 2.0 ghz still reporting 110c.

PS
Sisoft sandra was reporting 70c.
This story is true and thus i recomend not o/c over 25% of the stock speed
 

m25

Distinguished
May 23, 2006
2,363
0
19,780
hey this is my first post....

anyway i got myself a northwood 2.66 533fsb as well and i got it up to 3.2ghz sweet. only needed a 5% voltage increase and rean prime 95 for 8 hrs no probs. if u want to oc this cpu further might try goin into the bio and lowering the timings of ur ram coz that mite be the limiting factor. But i hit a wall when trying to do 3.3ghz with resets and lockups..

Oh im also using a thermalright xp 90 on a gigabyte GA-8S655FX-L as well
temps are: 26c idle
53c full load max
True that the Northwood is also a good OC-er, almost 600MHz, however not as good as the 3 year newer venice 3.2GHz is still 20% over 2.66 while the "normal" 2.6GHz of a 3000+ is almost 45% over the stock 1.8GHz. Stable up to 2.9GHz, it's been the overclockers' toy, we'll see what Conroe gives.
 

myou066

Distinguished
Aug 17, 2006
5
0
18,510
:?: lol mite jsut be a isolated incident? well anyway i wouldnt as well if im u, i just wanas squeeze every last drop out of my comp b4 i have to upgrade which im planning in dec on early next yr. U can oc it more than 25% its just that u need adequate cooling atleast if u want ur cpu to c the next day. btw how many ppl r stil using agp? as not many 478 socket motherboards have pci x..
 

m25

Distinguished
May 23, 2006
2,363
0
19,780
Well i have to be the king of (reported) temps :)
i have a p4 nortwood @ 2.0 ghz.Some time ago i was O/C him @ 2.6 but one day i noticed something !
cpu tmep 110c at the bios
i was like WTF!!!!
I gues i must have burned the thermistor some how :p
No o/c for him anymore.He runs at 2.0 ghz still reporting 110c.

PS
Sisoft sandra was reporting 70c.
This story is true and thus i recomend not o/c over 25% of the stock speed

Your temps might have been OK even on 2.66GHz;OC and check them by touching the heatsink (after shutdown). Your sensor should have been wrong because any prolonged state (1min+) of 110°C would heve already killed it, 100%.
I know that at least 2GHz celerons OC perfectly to 2.66 on stock cooling (FSB sent from 400 to 533MHz).
 

Grimmy

Splendid
Feb 20, 2006
4,431
0
22,780
Umm.. my P4 requires stock 1.525v. To run it stable OC, I need to raise it to 1.675v for 2.960. Anyhoo.. found out why it was not stable.

I don't have that problem with mine: from 1.80GHz to 2.05GHz it runs rock stable even undervolted (with 1.30V instead of the minimal 1.35). The Venice is a great core, almost brought a new era in OC-ing, where you can have 20-30%+ on stock. You do it and are 100% sure it won't do a scratch to your CPU.

Not really understanding why your explaining that to me. Okay, its nice that you can undervolt and still run stable at 1.8-2.05Ghz. But doesn't that have cool n quiet, which allows lower voltage?

I guess the only real benefit I'd get from that core, is if the power went out, and I was left to run on my UPS.
 

Grimmy

Splendid
Feb 20, 2006
4,431
0
22,780
hey this is my first post....

anyway i got myself a northwood 2.66 533fsb as well and i got it up to 3.2ghz sweet. only needed a 5% voltage increase and rean prime 95 for 8 hrs no probs. if u want to oc this cpu further might try goin into the bio and lowering the timings of ur ram coz that mite be the limiting factor. But i hit a wall when trying to do 3.3ghz with resets and lockups..

Oh im also using a thermalright xp 90 on a gigabyte GA-8S655FX-L as well
temps are: 26c idle
53c full load max

Interesting... So your in New Zealand (kinda cold there?)? Must have good room temps then. Well I knew my wall was at 3.2ghz. Just didn't want to fiddle around too much more, even though I thought I'd perhaps go to a P4 2.8c or 3.0c sooner or later.

I'm going to improve this box as much as I can to make it my future linux box, and perhaps build a conroe or AM2 machine for XP to Vista later on.
 

m25

Distinguished
May 23, 2006
2,363
0
19,780
I guess the only real benefit I'd get from that core, is if the power went out, and I was left to run on my UPS.

:lol: :lol: :lol: That was good! Talking seriously, it really makes few difference; the whole desktop PC is not optimised for that so you get almost no extra time.
However, I said it because I like knowing that the CPU runs cooler, near to ambient temperature. I just tried it @ 1.20V, hard rendering session: ultra stable and 5-6°Ccooler than 1.50V.
 

Grimmy

Splendid
Feb 20, 2006
4,431
0
22,780
I guess the only real benefit I'd get from that core, is if the power went out, and I was left to run on my UPS.

:lol: :lol: :lol: That was good! Talking seriously, it really makes few difference; the whole desktop PC is not optimised for that so you get almost no extra time.
However, I said it because I like knowing that the CPU runs cooler, near to ambient temperature. I just tried it @ 1.20V, hard rendering session: ultra stable and 5-6°Ccooler than 1.50V.

Heh, I was actually kinda serious about UPS time.

When I was talking about stable, I ment passing Prime95 torture tests.

How are you sure that it stays at 1.20 vcore under stress? I see your into rendering. Does it pass Prime95 (6 hours seesion) when undervolting it that low? I would guess it would.

I'm not exactly sure how low my voltage would go (bios only go to its lowest at 1.525v), if I underclock my P4 to 1.8 or 2.0 Ghz. But I'm pretty sure my temps would be lower then running stock 2.66 Ghz.

Edit: Doh.. the lowest Ghz I can go is.. 2Ghz. My mulitplier is locked at 20. :lol: