Why is my Opteron so hot?

[Ender17]

So I finally got my new Opteron 170 0530TPMW installed and my temps are way too high. At stock voltage and speed, this thing hit 53 C after 5 minutes of dual Prime95. Idle temp is 44 C.

Compared to my old 4400+, which never hit 50 C after 39 hours of dual Prime95 at 2.6GHz and 1.5v

Now I did change CPU coolers, but I thought I upgraded. Went from a XP-120 with Panaflo M1A to a Zalman 9500. Also noticed in my BIOS, it doesn't recognize the model number of my CPU, don't know if this means anything or not.

So, any suggestions before I put the X2 back in and sell this 170?


System specs:

Antec P180, 2 x 120mm Yate Loon fans, blowhole sealed off
PCP&C 510W PCI-E
ASUS A8N-SLI Premium
2 x 1GB Crucial Value
BFG 7800GT with NV Silencer
3 optical drives and 5 hard drives

[eBoy0]

Try updating bios, since its not detecting model of CPU temp sensor may be off.

[Ender17]

Try updating bios, since its not detecting model of CPU temp sensor may be off.

I was afraid you'd say that. I've had really bad luck with BIOS flashes. Killed my first version of this board that way.

Any tips?

[uOpt]

I have an Opteron 146 which is certainly returning rubbish from its temp sensor. It claims to be above 50 yada yada.

However, I have an independent temperature sensor directly at the joint between IHS and the Big Typhoon base plate, which is displaying normal temperatures.

Since the same thermal probe is very close to the temperature reported with other CPUs I figure this Opteron must have a bad sensor.

Either that or the IHS has insufficient contact.

However, on my powermeter I can see that this supposedly higher temperature Opteron doesn't take more power than other CPUs which report they run cooler. Since the heat is a direct result of power consumed I concluded it is a whacky sensor.

[eBoy0]

I was afraid you'd say that. I've had really bad luck with BIOS flashes. Killed my first version of this board that way.

Any tips?

Well doing it w/ a floppy is safest i'd guess, but i always just use winflash.

[deff_ins]

my 170 opteron is very hot too
it reach 60C at 2,9ghz priming on chilled water :o but that not the problem
dfi sli-d and expert mosfets get hot too - very hot... dont know how to cool everything down

[eBoy0]

60c on chilled water? Something is wrong... waterblock clog or bad contact?

[Absolute_0]

60c on chilled water? Something is wrong... waterblock clog or bad contact?

Or bad readings. 60C on chilled water is just about impossible, excepting the things you mentioned.

[dogsx2]

my 170 opteron is very hot too
it reach 60C at 2,9ghz priming on chilled water :o but that not the problem
dfi sli-d and expert mosfets get hot too - very hot... dont know how to cool everything down

How many volts is the 2.9 running on?

[tHe_jOkeR]

did u check the contact pressure?
maybe its the Zalman...

3200+wichester@1,4V reach 55°C after 1 hour prime
idle was 43°C ambiente about 25°C in my opinion the
zalman sucks...

than i realised that the bolts at the left an the right side
didn´t grap rightly.

at fixing this problem i just beat up the thread
so i dont know it jet if it was the contact pressure..

hth
greetz

[fullup3]

my opteron is hotter than any other i've heard of. It idles at about 44c at stock volts @2800, and loads at around 55c. At 3000 and 1.5v or so, it gets up near 60c. This is on cheap water. On the zalman 7700, it would get to 65c at load.
This is with the ihs removed also!
Not sure if these temps are correct or not, but they don't seam to be harming anything. My chip is still kickin' along nicely.

[ozzimark]

However, on my powermeter I can see that this supposedly higher temperature Opteron doesn't take more power than other CPUs which report they run cooler. Since the heat is a direct result of power consumed I concluded it is a whacky sensor.

i'm not sure if i follow.

if the contact is bad, the cpu will heat up more until it starts dissiating the same amount of energy through the heatspreader as a cpu with good contact would at lower temps.. at the same heat output. i'm definitly blaming the spreader contact in your case.

[manomanx2]

I'm thinking you should any or all of these things: 1. Let the thermal compound settle in; 2. Remount your 9500; 3. Bios flash. Also, have you your old XP-120 around? Try that and compare, maybe your cooler is faulty (although I have never heard of that being the problem).

[Ender17]

I survived a BIOS upgrade
now my CPU shows up in the BIOS, but no effect on temps
just hit 56 C after a few minutes of dual Prime95

gonna swap the X2 in later

[Ender17]

I checked the HSF and it did sem a bit loose, but tightening it didn't help anything
with ClockGen, I went up to 2.4GHz, stock voltage, before temps hit 60 C under dual Prime95

so I turned it off and put my X2 back in
currently at 48 C, stock everything, dual Prime95

here are some pics of my application incase I'm doing anything wrong

this is what it looked like when I took the Opteron out



and my dab before putting the HSF on the X2

[MaxxxRacer]

more proof that the grain of rice theory is a load or horse crap.

[Ender17]

more proof that the grain of rice theory is a load or horse crap.

please explain

[mongoled]

more proof that the grain of rice theory is a load or horse crap.

In saying this, are you insinuating that if you put more paste you will get better temperatures? If you are then for me you are not correct. The amount of paste only needs to be enuff to create a thin layer between the IHS/Core and heatsink/block anything more is simply wasted, this is simple thermodynamics, not rocket science. Any results showing otherwise are because of other circumstances, this is my opinion, and its not open for discussion, lol

Ender17, to be really sure you need to lap both the IHS and Zalman 9500, seeing tht the Zalman 9500 performs fine on your X2, then a lapping of the Opty would be needed, problem here is tht u loose your warrenty . I have found tht the A64 CPU's I have had, are not getting pressure from the heatsink in the correct place i.e. in the center of IHS because of shape of IHS and heatsink. Just imagine wht happens if the the corners of the IHS are higher than the center if IHS, the heatsink will be applying pressure to the corners, creating a 'vaccum' of pressure in the center. I think this is the reason we find many IHS with corners sanded down from the factory.

Ofcourse you may do everything you can to decrease temps, but in reality your chip is a not one, i think it all comes down to how much u want to know, wht the cause is

[TEDY]

I installed my first Opteron 146 now FPBW 206 batch

At load around 43-45°c default 200*10......freshly installed...think AS5 needs to work some time now

[demonbane23]

my opt 170 runs super cool and the reading is borked the other way,,, I am at 17c idle and 30c load with a 2.6ghz overclock,,,,

but when I ask or browse at DFI street they say its my fault for using an unsupported CPU,,,, ehh whatever im stable and the watercooling is doing its job,,,,

I cant stand reading posts at DFI street all the users just bash anyone who is having a problem, very rude and ignorant crowd it seems,,,,

[situman]

my 170 reached 57c on my Sapphire when running two instances of cpuburn. at 2.7ghz at stock volts. Promptly took the IHS off and temps dropped to 44c to the 47c range. These things run hot. My radiator is cold to the touch though. So go figure.

[uOpt]

i'm not sure if i follow.

if the contact is bad, the cpu will heat up more until it starts dissiating the same amount of energy through the heatspreader as a cpu with good contact would at lower temps.. at the same heat output. i'm definitly blaming the spreader contact in your case.

You are right, I wasn't thinking straight.

Maybe this is a good candidate chip to pop my first IHS.

[periquitos]

170 owners , could you please report your tmaxcase on the tmcase thread?

Mine is as 170 0530TPMW and is very hot. Hot chips with a good cooling appears to overclock better, is this right?

[periquitos]

this thread:
http://www.xtremesystems.org/forums/...ad.php?t=71044

[Übergeek]

I checked the HSF and it did sem a bit loose, but tightening it didn't help anything
with ClockGen, I went up to 2.4GHz, stock voltage, before temps hit 60 C under dual Prime95

so I turned it off and put my X2 back in
currently at 48 C, stock everything, dual Prime95

here are some pics of my application incase I'm doing anything wrong

this is what it looked like when I took the Opteron out



and my dab before putting the HSF on the X2

That is WAY too much tim!

You should use like a fourth of that, tim is mearely for filling up tiny little holes in the metal, you put on so much tim that it's isolationg, see the bottom of your heatspreader it's all bubbly it should be way thinner than that!