How hot is too hot for a Prescott SL7E2?

[Duvie]

Those are the limits of what Intel has viewed as acceptable opreting temps...Meaning you can run outside of that range but it may effect the longevity of the cpu. Acanthus is right...After that point it should start to throttle....Now the real question is where is that temp taken from in the cpu and how far off are you with your temp readings....

When you run WMD go ahead and start throttlewatch and place it in the background it will record...Temps will lie but if the cpu throttles lower mhz speed to compensate high temps then there is a temp issue...The fact you can play and only rais etemp 2c makes me think something is BS here....With obviously sub par cooling now and the fact that NOOOOO current HSF even water block would only have a2c swing from idle to load we are basically looking at bogus numbers...

Just run the flippin throttlewatch and run you WMD, prime95 (same time).....If you are not running 2 cpu intensive apps you are kidding yourself again cause it wont fully load the cpu and give you the true hottest temp. You need HT or 2nd virtual cpu to be loaded fully as well....

 

[computer]

Originally posted by: Duvie
Those are the limits of what Intel has viewed as acceptable opreting temps...Meaning you can run outside of that range but it may effect the longevity of the cpu. Acanthus is right...After that point it should start to throttle....Now the real question is where is that temp taken from in the cpu and how far off are you with your temp readings....

When you run WMD go ahead and start throttlewatch and place it in the background it will record...Temps will lie but if the cpu throttles lower mhz speed to compensate high temps then there is a temp issue...The fact you can play and only rais etemp 2c makes me think something is BS here....With obviously sub par cooling now and the fact that NOOOOO current HSF even water block would only have a2c swing from idle to load we are basically looking at bogus numbers...

Just run the flippin throttlewatch and run you WMD, prime95 (same time).....If you are not running 2 cpu intensive apps you are kidding yourself again cause it wont fully load the cpu and give you the true hottest temp. You need HT or 2nd virtual cpu to be loaded fully as well....

Thanks for the explanation.
Now the real question is where is that temp taken from in the cpu and how far off are you with your temp readings....
I don't know if you're asking me that, or rhetorically asking Intel that. 😉 As I indicated before, the temps I'm stating are from either in the BIOS (those of course are the high readings); or the sensor I have taped to the underside of the mobo where the CPU socket is--which as been about 96°-104°F. I previously had one sensor under the HS touching the CPU's IHS, and it never got anywhere near even 110°F.

Again, this is not a complete PC. The customer is only getting the mobo, CPU, memory, and PS & case (and of course a video card is on it now). So, I'd have to hook up a HD, install XP, then install other software, unless this "Throttle Watch" can be run from a floppy. WMD only runs from a floppy in pseudo DOS, just like MemTest.

Are you saying you think the mobo's thermal diode/sensor is screwed up, inaccurate?

 

[computer]

Originally posted by: Duvie
...If you are not running 2 cpu intensive apps you are kidding yourself again cause it wont fully load the cpu and give you the true hottest temp. You need HT or 2nd virtual cpu to be loaded fully as well....

I forgot to mention this is not a HT enabled CPU. HT Prescotts are only 800mhz bus I believe.

 

[Duvie]

OK...i didn't catch the HT thing...that helps you somewat as HT runs even hotter.....

Until you get into a windows environment which I recommend I would seriously doubt the real satbility of this system...Bios temps are usually quite bogus for some reason...My MSI bios temp like never changes no matter if I am 1.8ghz to 2.66ghz with lots of core it moves like 1-2c idle.....That is BUllsh^t!!! I dont trust Bios temps alone...You need to run throttlewatch cause if WMD only raised temps by 2c then it didn't either load it enough, or is showing how false it is.....

The temp sensor under the HS between it and the IHS is the best that one can test but the temperature is likely being measured inside the core that effects throttling. I dont really know if you can really get anything accurate within 10c (18f) really......


If you are also testing this outside of a case and the temps may be in question then placing this inside a cramp case with possibly not enough fans or ineffective flow will quickly exasperate this and the swing from idle to load will grow...Again 2c swing from idle to load is impossible with any cuurent HS material wise, thermal paste, and cfm fans.....Even water is going to have 6-10c swing.....


EDIT: I would say with a reported 129f or 54c and by all means normal HSF cooling (38cfm is no great feat and running 60mm fans just makes the noise unbearable in my opinion. I have a YStech 80mm that did 48cfm at 3800rpm....I also had a moded heatsink top to place a 92mm fan (panaflo) that did 52.1cfm at like 2700rpm....)....Based on what I have said a 2.8ghz prescott (hotter then comparable 2.8bs or c's) with default vcore should still have a idle to load swing of at least 18-22c with adequate case cooling...therefore that 54c will be to mid 70's and would definitely throttle if it was within 6c of accuracy....

 

[imported_mystro]

Originally posted by: clarkey01
First I really dont understand why anyone who wants to go the Intel route buys prescott. What the hells wrong with Northwood ? Yes I know it's 400 Mhz behind the flagship (3.4 Ghz Vs 3.8 Ghz) But how much more you need ?

Did you do your homework on this CPU before you bought it ? its slightly slower then northwood and runs hotter.

As for the heatsink issues you seem to have, I'll leave that to the more experianced members of the board.

The Northwood is getting much harder to find especially new retail/oem and always gets the premium price (and worth it natually)..rumor has it Intel grabbed all of them 🙂

 

[computer]

Originally posted by: Duvie
OK...i didn't catch the HT thing...that helps you somewat as HT runs even hotter.....

Until you get into a windows environment which I recommend I would seriously doubt the real satbility of this system...Bios temps are usually quite bogus for some reason...My MSI bios temp like never changes no matter if I am 1.8ghz to 2.66ghz with lots of core it moves like 1-2c idle.....That is BUllsh^t!!! I dont trust Bios temps alone...You need to run throttlewatch cause if WMD only raised temps by 2c then it didn't either load it enough, or is showing how false it is.....

The temp sensor under the HS between it and the IHS is the best that one can test but the temperature is likely being measured inside the core that effects throttling. I dont really know if you can really get anything accurate within 10c (18f) really......


If you are also testing this outside of a case and the temps may be in question then placing this inside a cramp case with possibly not enough fans or ineffective flow will quickly exasperate this and the swing from idle to load will grow...Again 2c swing from idle to load is impossible with any cuurent HS material wise, thermal paste, and cfm fans.....Even water is going to have 6-10c swing.....


EDIT: I would say with a reported 129f or 54c and by all means normal HSF cooling (38cfm is no great feat and running 60mm fans just makes the noise unbearable in my opinion. I have a YStech 80mm that did 48cfm at 3800rpm....I also had a moded heatsink top to place a 92mm fan (panaflo) that did 52.1cfm at like 2700rpm....)....Based on what I have said a 2.8ghz prescott (hotter then comparable 2.8bs or c's) with default vcore should still have a idle to load swing of at least 18-22c with adequate case cooling...therefore that 54c will be to mid 70's and would definitely throttle if it was within 6c of accuracy....

I'll be putting at least 3 fans in the case + a 3-fan 5.25" bay cooler. Right now there's a 105cfm 120mm front fan, and two 80mm rear fans + the bay cooler.

Well, I found out the problem. This mofo (either the CPU or the HSF unit) is WARPED! I removed the HS to put the final TIM on it and place the sensor, and part of the "contact area" of the CPU to HS was BARE! An open spot totally clean of TIM! So, the two are not even making contact. I put the retail Intel HS on it to check contact and it is making total contact. But this go*damn POS Thermaltake A4012-2 is not, so it is what is warped. That makes two f**king HSF units I've blew my money on!!!! Now I'll learn the hard way to stick with the retail Intel HSF unit. This would explain the poor showing of the A4012 unit since it's not making total contact, and would also show just how great AS5 is. If it can decrease temps by 6° with only about 50% contact, that stuff must be awesome. So this now begs the question: just how thick of a layer of AS5 can be used before it becomes a thermal insulator?

 

[JBird7986]

Does anybody else here find it sad that we need to turn on our computers and submerge them in a vat of liquid nitrogen in order to stop them from frying at idle nowadays? At any rate, after reading this thread I think I finally understand why Intel pulled the plug on the 4GHz chip now. I had heard they were melting, but now I finally understand how that is possible.

 

[pplapeu]

Originally posted by: computer
It just never ends. I got the new Ttake A4012-2 in and believe it or not, it's now running HOTTER!! 131°F!!!! :brokenheart: :roll: :disgust: :| The board is also shutting down and I have no idea why! :roll: :disgust: :| CPU's shutdown temp is disabled, and the warning temp is set to 151°F. Three times now it has shut completely off all by itself while I was watching the temp in the BIOS. I removed the 80mm fan that came with it, (which was no surprise also overrated at 32+ CFM). It's actually 30cfm. I put a 60mm 38cfm fan on it and temps got back down to below 129°, and it's still shutting off with absolutely no warning. (I used the TIM that came on the HS, it's not a pad. Ttake uses a patch of white Dow Corning TIM).

I noticed where the fan's steel mount connects to the copper HS with screws was loose. I tightened them before I put it on the CPU. Now I'm wondering if they were supposed to be loose to allow for thermal expansion of the copper. 😕 I'm not sure what this would have to do with it shutting down though. Unless the screws being tight not allowing the copper to expand freely is putting too much pressure on the copper, therefore too much pressure on the CPU. ????

888888888888888888888888

Heat is exciting. I used to run my 3.2E in the upper 60s to 70C Celsius playing new 3D games and it played real well and never shut down. I even bumped up the alarms to 72C and kept on at it without the beeper going off. Someone in the Asian tropics said that Intel is happily letting the users run up to 75C with no problems.

Anyway I went to home made H2O that costs as much as the CPU as I know I can use the parts for the water in an upgrade whenever. Just ball park I lost 10c to 15c with water and the games seem a bit better.

I was worried about the hot 6800 GPU i got so I touched the card right behind the GPU and felt very little heat, not even electric blanker warm heat.

Why do you not put your finger as close to the CPU as you can and see if you feel something. Like ohers said the MOBO readout might be over compensated for safety or whatever.

 

[computer]

Originally posted by: pplapeu

Originally posted by: computer
It just never ends. I got the new Ttake A4012-2 in and believe it or not, it's now running HOTTER!! 131°F!!!! :brokenheart: :roll: :disgust: :| The board is also shutting down and I have no idea why! :roll: :disgust: :| CPU's shutdown temp is disabled, and the warning temp is set to 151°F. Three times now it has shut completely off all by itself while I was watching the temp in the BIOS. I removed the 80mm fan that came with it, (which was no surprise also overrated at 32+ CFM). It's actually 30cfm. I put a 60mm 38cfm fan on it and temps got back down to below 129°, and it's still shutting off with absolutely no warning. (I used the TIM that came on the HS, it's not a pad. Ttake uses a patch of white Dow Corning TIM).

I noticed where the fan's steel mount connects to the copper HS with screws was loose. I tightened them before I put it on the CPU. Now I'm wondering if they were supposed to be loose to allow for thermal expansion of the copper. 😕 I'm not sure what this would have to do with it shutting down though. Unless the screws being tight not allowing the copper to expand freely is putting too much pressure on the copper, therefore too much pressure on the CPU. ????

888888888888888888888888

Heat is exciting. I used to run my 3.2E in the upper 60s to 70C Celsius playing new 3D games and it played real well and never shut down. I even bumped up the alarms to 72C and kept on at it without the beeper going off. Someone in the Asian tropics said that Intel is happily letting the users run up to 75C with no problems.

Anyway I went to home made H2O that costs as much as the CPU as I know I can use the parts for the water in an upgrade whenever. Just ball park I lost 10c to 15c with water and the games seem a bit better.

I was worried about the hot 6800 GPU i got so I touched the card right behind the GPU and felt very little heat, not even electric blanker warm heat.

Why do you not put your finger as close to the CPU as you can and see if you feel something. Like ohers said the MOBO readout might be over compensated for safety or whatever.

I've posted several things since then. If you check you'll see where I mentioned the bottom of the HS was dented, concave and it was only making about 50% contact. The incredible thing is NewEgg since a new one overnight for free, it's flat, and it's still running 120-123°F!!!!!! (123F = 50C). Which I believe is where it was with the dented HS and AS5. :Q The thermal sensor I put under the HS as close to the CPU's IHS as I could get it reads about 90°F. 😎 The underside of the mobo opposite the CPU socket was never hot to the touch, just warm. Also rather strange is that temps remained the same after putting it in the case (front 120mm 105cfm fan, 2 - 80mm rear 32cfm & 39cfm, and one 5.25" bay cooler w/3 - 40mm fans about 9cfm each. PS has dual fans--one sucking up from the bottom and the typical one in the rear exhaust). AS5 needs a 200 hour "break in" according to AS, and they say temps will drop 2-5°C after that.

Apparently the mobo's sensor is spastic. :evil: :frown:

 

[MrDave]

What Motherboard is this on?
If it is an Abit, then 54Cto 60C in Bios is perfectly Normal. Bios CPU temps are the same as Under Full load, for the most part, because none of the cooling featurees of the CPU are enabled, as when you are booted into Windows or Linux.

I'd like to make a suggestion on a cooler. I understand the person doesn't want to spend money on one, but a "Heatpipe" cooler WILL drop the cpu temps by 10 to 13 degrees celcius!!

I use the Thermaltake Silent Tower Heatpipe with aluminum fins, a 92mm fan with sideflow intakes and it is SILENT, as the name suggests!!!

Also I run an Abit IC7 motherboard, I have a thermal sensor installed at the cpu heat spreader and the contact surface of my cooler which reads 34C Idle and 46C load.
However, my MB sensors read 14C higher then those numbers!!! Then add on not using a Heatpipe cooler, another 10C to 13C and the numbers are where you are reporting them!

Abit motherboards report your temperature over 10C too hot, this is a known fact!!!

Also motherboards that do not have a temp sensor software package available, such as the Soyo Dragon Plus v1.0 for Prescott CPU's, do NOT give an acurate reading with applications such as "Hardware Monitor", because the sensors, although available on the motherboard, have not been engineered and aligned for propper operation with any temp sensing software solution.

I tend to trust a good quality external tempature sensor any day over my Abit MB sensors!
And I kid you not, my CPU temps dropped 13C with the Copper Heatpipe solution !!!

My rig-
Abit IC7 mb with 2 Gbyte 2.3.2.5T1 Dual channel DDR
Intel Prescott SL7E6 3.4 Ghz CPU with ThermalTake Silent tower heatpipe cooler
ATI All In Wonder 9600XT

 

[computer]

This was on a QDI P2PE/800 PRO, Intel 845 chipset. Best I was ever able to get it was 123°F as per the BIOS. The temp sensor I put under the HSF unit up against the edge of the CPU IHS never got above 100°F. I used AS5, and the HSF unit I ended up getting was a ThermalTake TR2-M12 (#A4012-02).

 

[MrDave]

Originally posted by: computer
This was on a QDI P2PE/800 PRO, Intel 845 chipset. Best I was ever able to get it was 123°F as per the BIOS. The temp sensor I put under the HSF unit up against the edge of the CPU IHS never got above 100°F. I used AS5, and the HSF unit I ended up getting was a ThermalTake TR2-M12 (#A4012-02).

QDI was never ranked as a top manufacturer however their boards do work. The website was so slow, so I gave up waiting and looking into that model MB to see if it had a supporting Hardware monitor application.

I would tend to believe your external temp sensor,which you placed in several key locations, as close to the heat source as possible over the MB sensor in this case.

The 100'F sounds fine. I wouldn't worry about that system. However, I would recommend upgrading to some form of Copper with Heatpipe cooling solution, just to keep this cooling down to a respectable temperature.

Also check out this thread, Here, which talks about the very warm running Prescott series CPU's and asks for people to post their temps and experiences. Here you can gain a little insite on other people temps, cooling solutions and opinions..