Prescott core original... too hot

I

Imcompa

Weaksauce
Joined
Aug 10, 2004
Messages
109
I bought one of the original prescott core 3.0E pent 4 way back when it came out and it has always burned hot as fuck, cooling isn't an issue and it's always had an internal temp of like 70*C when running at 100%, it crashes in under 5 minutes at 100%

I just want to get a new cpu that fits the same computer but isn't as hot, it can be slower or not, any ideas?

my mobo is a an Abit IS7-E2 (i865-ICH5)
 
http://www.zipzoomfly.com/jsp/ProductDetail.jsp?ProductCode=80663&affiliate=froogle

Although, I don't understand your post. You say cooling isn't an issue, yet "it's always had an internal temp of like 70*C when running at 100%, it crashes in under 5 minutes at 100%". Anyway, the chip I linked to is a good deal, if you're looking for a cooling running cpu for roughly the same performance...BUT, I'd invest in a different hsf or check your seating before shelling out the loot for a new chip. 3.0/3.2 Prescott's are great o'clockers!

Good luck!
 
By cooling isn't an issue i meant it was the processor that burned too hot, I have a decent heatsink and thermal paste, other processors ran fine in the case.
 
Imcompa said:
By cooling isn't an issue i meant it was the processor that burned too hot, I have a decent heatsink and thermal paste, other processors ran fine in the case.
Click to expand...

Well, what other processors did you run in that motherboard? I have the same mobo (minus E2) and when I went from P4 2.8@3.6GHz Northwood to P4 3.0@3.0GHz Prescott, my temps went up by about 7~9 deg C at default with an older Zalman heatsink. Had to get a new (bigger) heatsink to compensate for the temp difference and to be able to overclock. But if you're convinced it's the CPU itself that's running way too hot, get a hold of Intel and request an exchange.

Mike
 
If your processor burns "hot as fuck", then you're cooling is not fine...It IS possible to get S478 Prescotts running cool but you're not going to do it with something like a Volcano 7.

We need to know what cooling you do have.
 
You might want to reseat your heatsink. I've got a 3.8 that doesnt get over 60c at full load with a crap generic heatsink (this pos)
 
I have a Celeron D 2.4@3.8 that doesn't break 60C with a Zalman 7700CU on the slowest speed. The stock Intel coolers are actually pretty decent coolers, they won't compare with high end units but they can hold their own in the midrange. Intel has definetly gotten better with their coolers since the early P3/P4 days.
 
acascianelli said:
Intel has definetly gotten better with their coolers since the early P3/P4 days.
Click to expand...
They didnt really have a choice, theyve got the power of a light bulb coming out of a tiny silicon wafer, lol
 
lithium726 said:
...theyve got the power of a light bulb coming out of a tiny silicon wafer, lol
Click to expand...

Heh, never thought of it like that. Scary that its true.
 
I'm not sure why someone would post that their CPU is running too hot, post their CPU type, motherboard make and model, assert that it isn't a cooling problem (wtf?), and then (more than once) fail to list the make/model of the heatsink being used. Sounds like denial. ;)

I recently put a 3.2GHz Presscott into a user's Dell GX-270 that had a crappy cooler only meant for Northwoods only up to 3.0Ghz (passive aluminum heatsink with no direct fan). It ran too hot to touch. I eventually exchanged it for a Northwood 3.2GHz (ironically, from ZZF just as that person linked above). It's great that ZZF still has these CPUs on-hand.

But, really, with an IS7, you have the flexibility to mount just about any cooler (given case constraints). . . and you *can* get a Presscott down to a decent temperature with a decent cooler (XI-120 or Ninja?) at about a quarter of the cost of a Northwood CPU.

BTW, I had to use zip-ties to strap a fan onto the Dell heatsink even with the Northwood CPU. The next person who opens that case will think I was on crack. But after four days struggling with their proprietary motherboard layout (stock cooler retension mechanism *and* custom ones won't fit due to capacitors encrouching on CPU area), I really decided I didn't have a choice. It's on securely and I tested for throttling. . . it's all good!
 
sorry guys, thanks for all the help, I have the zalman all copper heatsink, i think it's this one CNPS7000B-Cu but i don't remember, it was like 40$ a year ago, I use the generic paste, and when that didn't work, i tried a friends silver(silver whatever) thermal paste, and that didn't work, so he tried applying and them temp remained the same. I'm working on some screenshots and such to show the temperature idle, when running at ~50% and when at ~100%
 
mmk I couldn't get a 100% because in under 20 seconds the computer shutdown, the heat started spiking and then a hardboot, no error.

idle:
idle.jpg


50% after 5 minutes:
50-numbers.jpg

50-chart.jpg


50% after 10 minutes:
50-2-numbers.jpg

50-2-chart.jpg


back to idle:
backtoidlenumbers.jpg

backtoidle.jpg
 
Here is my speedfan screenshot, what makes noise in yours is the power supply voltages, looks weird to me

speedfan2ir.png
 
I dont think that is your CPU, Im sure its either the amount of paste you put on or the heat sink is not up to par for that CPU. 20C increase from Idle to full load on air cooling is pretty common. Id recomend get your hands on a Big Typhoon or something in its class. Should drop your temps way down. My 820 and 830 are small ovens when they are loaded and my old Typhoon kept them quite cool. I was able to get a 950mhz overclock with a load temp of 55C.

Heres the proper way to apply thermal paste on a single core CPU( Arctic Silver 5 was used)

3464


3462


If it is applied as above, then it should look similar to this when you remove the heat sink later.
3463


Just needs a dab of paste, most people say spread it around but the heatsink will spread it alot better and wont leave air pockets between the paste and the heat sink.
 
Cool, thanks for the added info. That cooler should be decent (though not really considered high-end by today's standards). Have you tried running "Throttlewatch" to check for CPU throttling as it increases in temp? To be honest, I'm surprised that it's freezing/crashing at 70c at stock speeds. While very hot, I've seen hotter! :p

Edit:

Woops, you just posted screenies (had this window open for a while). . .

Okay, the shutdown behavior is probably due to your BIOS. What BIOS version is your IS7 running? It has a temp shutdown threshold that is probably being exceeded. Is it possible that you are running an older BIOS revision that is applying (by default) Northwood-era temp thresholds towards (hotter) Prescott CPUs?

Regardless, I'd go in there and disable (or raise) the shutdown threshold. I used to have an IS7 (still do, actually, but it's in my "Tivo" box and I rarely muck with it anymore, it just works), so I'm almost positive that this setting exists in that board's BIOS. Then, I'd use "throttlewatch" to see if your CPU is actually throttling when under load. And, of course, look for a newer BIOS for your IS7.

Of course, you chould also make absolutely sure that your CPU is making good contact with your heatsink (as the above poster mentions), and possibly consider a new cooling unit.
 
Oh, and one other thing! The IS7 BIOS was notorious for over-reporting CPU temps until one of the later BIOS revisions! So, your temps may not actually be as high as reported. Check out newer BIOS availability!

Wow, all these Abit IS7 memories are gradually coming back to me. . . :)
 
Thanks for all the help, I remember when i first got it it wouldn't even start until i updated the bios due to the early bios having like 15*C too high of a reading, I'm going to try reapplying thermal paste and watching this throttle watch. Thanks for all of the help.
 
I wanted to add one more thing... When ever I try and rip music using windows media player, it spikes my cpu and in under 20 seconds the computer just shuts down, I tried disconnecting my dvd drive to make sure that it wasn't a power thing, but when you said my power was looking weird i thought I might mention that.
 
articsilver5.jpg

I bought some artic silver 5 today and cleaned it up and reapplied it. I believe that is a slight decrease. :\

50% after 10 minutes
50articnumbers.jpg

50articchart.jpg
 
I watched throttlewatch and the voltage chart and the temperature chart all while maxing out the cpu and the computer just crashes at around 58*C after about a minute. Are there any other monitoring programs that I can use to monitor different things, maybe some that write as they record so i can see the last things that happen before a shutdown
 
Well that helped a little bit, looks like the HS needs to be replaced with something better.

Give this a gander. 1 of the best air coolers on the market.

Here is another nice air cooler. This is makeing a name for its self in highend air overclockers. It is quite huge, but seems to do an awsome job.

WIth either of those you will probably drop 10-20C off your load temps, but 1 thing to remember your room temp has alot to do with the PC temps. SO if you have a warm room then your going to have a toasty CPU.

I watched throttlewatch and the voltage chart and the temperature chart all while maxing out the cpu and the computer just crashes at around 58*C after about a minute. Are there any other monitoring programs that I can use to monitor different things, maybe some that write as they record so i can see the last things that happen before a shutdown
Click to expand...

Well if it keeps crashing with lower temps you have either ran the CPU to hot to long, or your ram is getting ready to puke. I guess you could try some differant ram for awhile, maybe try a differant PSU. None of these will affect your CPU temps but maybe its not your CPU that is crashing.

Download this program. and save it to a bootable floppy drive, or burn the image to a CD-R as an image, setup your floppy or CD rom to boot first (which ever you choose to use) and let the program run at least 2 times. If it doesnt have any errors then its not your ram.
 
i was wondering about the ram, I'll run that and see, thanks again :)

right now i'm runing a CPU Stability test and watching it alls.

edit: during this stability test it doesn't shut down at all, it's running in the low 60*C's and it's running around 100% once it finishes i will run the ram test and that could very well be the issue.
 
lithium726 said:
They didnt really have a choice, theyve got the power of a light bulb coming out of a tiny silicon wafer, lol
Click to expand...


Actually the S478 Prescotts put out more heat per cubic inch than a nuclear reactor.
 
I've gotta say, your computer should not be shutting down or crashing with a Prescott in the mid-60s. At least, not without some underlying hardware fault.

I'd start suspecting:

1. RAM
2. CPU
3. Motherboard
4. Power Supply

I was skeptical at first that the power supply could be the culprit, but if your system is shutting down and/or behaving this erratically, it could very well be the problem. Regardless, however, I'd still get a better cooler on that Prescott. Running it cooler can't hurt, might help.

Best,

H
 
Alright so as we speak I'm running at CPU test and the program has been running the CPU at 100%~ for the last 20 minutes, the temp spiked at 62 and its at around 59*C now. I'm going to now move on to testing ram, and if it isn't ram, which i'm starting to suspect i'll move to finding a friend who will let me borrow a kgp power supply. thanks hurin and adeo
 
Yup sounds like something else if you are having no problems with the CPU at this point. Not sure if you picked up what I was saying in the other post, but you need to boot into DOS mode to run that Memtest. Im sure you figured that out but just thought Id specify. Good luck.
 
Hurin said:
Oh, and one other thing! The IS7 BIOS was notorious for over-reporting CPU temps until one of the later BIOS revisions! So, your temps may not actually be as high as reported. Check out newer BIOS availability!

Wow, all these Abit IS7 memories are gradually coming back to me. . . :)
Click to expand...

Oh god did the IS7 report high. It is a wonderful mobo (my favorite i865PE board) but the bios just plain lied about temperatures. A 2.8 northwood on stock cooler idling at 57C (one of my experiences)? I call BS. And even if those temps are correct they're not high enough to cause any real system problems. Temps up near 70C are pretty hot but they're pretty typical of the early 478 prescotts (luckily the lga775 prescotts are boatloads better). A friend of mine has one of the original dell inspiron xps laptops (yuck) with a 3.4 pressie in it...it routinely runs up in the high 60's and low 70's under load, no stability issues.

I say check out the cpu with throttlewatch...that should give the final call. No throttling = no temp problems. But I really think something else is at play here, like RAM or powersupply. How do the rails on your powersupply look? And how do they change when the system goes under load? The PSU is my prime candidate right now.
 
I am currently running that test on the memory thing, I'm not sure if it just continuously loops or if it's just a really long test but I already have 2 errors:

tst
5
5

pass
13
13

failing address
003e868440 - 1000.5mb
003f7e03a0 - 1015.0mb

good
00200000
00200000

bad
00280000
00280000

err-bits
00080000
00080000

count
1
1



(I'm 9 hours into it, do I just let it go until it says finished or something?
 
Damn, I completely forgot about the IS7/IC7 temperature reporting problem. I have both an IS7 and IC7-G and both of them report CPU temps around 8-12C higher than actual. There was a bios update that was supposed to recalibrate the monitor but IIRC it only worked with Prescott CPU's and it didn't fix the problem completely.

However, even if its reporting 70C it may be at or above 60C, which is still kinda warm. My Celeron D 2.4@3.8 maxes out at 62-63C which is actually around 55C.
 
Imcompa said:
I am currently running that test on the memory thing, I'm not sure if it just continuously loops or if it's just a really long test but I already have 2 errors:

tst
5
5

pass
13
13

failing address
003e868440 - 1000.5mb
003f7e03a0 - 1015.0mb

good
00200000
00200000

bad
00280000
00280000

err-bits
00080000
00080000

count
1
1



(I'm 9 hours into it, do I just let it go until it says finished or something?
Click to expand...


That means your ram is bad, more than likely. :\
 
I'm not sure if it just continuously loops or if it's just a really long test
Click to expand...
it continuously loops. from the results, it looks like you have a bad stick(or possibly a bad slot). try rerunning memtest with each stick seperately, see which one is reporting errors. if both sticks come back with no errors when run seperately, run the test with 1 stick in each different slot to see if you have a bad slot.

i was having restart troubles with my old k7n2 board, i replaced the ram repeatedly, but was still getting mem test errors, turned out the middle slot was faulty. 1 rma later, everything was hunky dorry ;)
 
Just a quick memtest tip. . . on the right-hand side, you should see a column labelled "pass." That's how many times memtest has run through the series of tests. I'd usually give it 2-3 passes without errors before giving the RAM a clean bill of health. More if overclocking.

Just mentioning it so you don't waste too much time waiting for it to stop on its own or wondering when "enough is enough."

Though not 100% indicative, Memtest reporting errors is a very good indication of RAM problems (thank you Cpt Obvious!). So, as others have said, take out one stick at a time and re-test.

I think you're closing in on the problem. If it's RAM, you're lucky. Easy to replace and not terribly expensive. :D
 
I switched the sticks and I'm running the test again. what bothers me is those two errors only occured once during 15~ hours of the test.
 
Not a single error after switching the sticks and running the test for 5 hours, i think it ran like 21 times, so now I borrow a PSU and see if I can fix this sometimes random shutting down/shutting down 20 seconds after attempting to rip music from a cd. thanks for all the help, I'll update the results of my next test.
 
So, you can usually get it crash/shut down at will when you max out the CPU in a certain way, correct? I'd take out one stick and try to replicate that crash. Then, if it doesn't crash, swap the sticks.

Of course, if it crashes as expected with one of the sticks, that's a great indicator. If not, it doesn't necessarily rule out RAM since it could just be the "loss" of dual-channel mode that is making the system more stable. So, crashing with only one stick would "prove" a stick bad. But no crashes doesn't necessarily prove both sticks to be not bad (impossible to prove a negative).
 
Pfft, Prescott too hot? Tell me about it.
The last week I was at work I had to replace the HSF on three different systems. Some of them would hit 85C before getting into windows.
 
You must log in or register to reply here.
Back
Top