Phenom II temperatures

[amphibem]

Hi there,

As you can see by my join date I have been around a while - but been pretty inactive for the last few years. Anyway I just upgraded by system with a Phenom II x4 955 Black Edition and ATi HD5770: still haven't gotten over how much grunt you can buy for what used to be entry-level dollars!

So given I was going from a 65W CPU to a 125W beast I expect a jump in temps, but not 90degC! That's what I currently get maxing out 4 cores folding, with a big Silverstone NT06 aftermarket cooler with 120mm fan on top (@12V) and 2 other 120mm fans in the system. Everything I read puts a max temp a 62degC so this cannot be good, once I have sent this will shut down and re-apply Artic Silver.

I get the same readings with Core Temp, Overdrive and in the BIOS so not bogus, any thoughts on what going wrong (other than poor contact, as above going to work on that)? For that matter why hasn't the machine shut down?

Thoughts appreciated, will check back in with new thermal paste coating applied.

[morphine]

The thermal paste probably isn't the problem, but it sure sounds as if the heatsink isn't sitting right on the CPU.

[Flying Fox]

If it is a Gigabyte motherboard check in CPU-Z to see if they are overvolting at the same time.

[amphibem]

So I took the heat sink off and found a couple of problems: the thermal paste was in somewhat of a lump (not spread over the full chip) and some of the nuts on the mount were loose. So I did a proper job of applying AS, fixed the mount and restarted. Now I get around 40degC at idle, still breaking 60 under load.

morphine it is an MSI motherboard, VID shows at 1.35V on all 4 cores. Currently sitting at 63degC, 30% CPU and fairly warm here (probably 25degC+ during the day).

My current cooler is a NT06, which really should be adequate for the job. It is noticeable that the metal of the cooler, and the air being pushed out, is getting pretty warm and was almost too hot to touch when the CPU was reading 90degC. This would lead me to assume the heat is being transferred off the CPU.

So pretty much at a loss. Will remount the standard backplate when I get a chance and try the OEM cooler.

[morphine]

I have to ask: why is it a loss? 60ºC under load with warm ambient temperature isn't anything to sneeze at. You know that desktop CPUs are usually rated, if I recall correctly, to 85-90ºC, right?

[insulin_junkie72]

I have to ask: why is it a loss? 60ºC under load with warm ambient temperature isn't anything to sneeze at. You know that desktop CPUs are usually rated, if I recall correctly, to 85-90ºC, right?

As mentioned, according to AMD, max temp on a 955 is 62 deg. C., so you don't have quite the thermal headroom of an Intel.

[amphibem]

I have to ask: why is it a loss? 60ºC under load with warm ambient temperature isn't anything to sneeze at. You know that desktop CPUs are usually rated, if I recall correctly, to 85-90ºC, right?

Well as per the next comment my understanding was Phenom's are rated to 62deg, hence my concern. Otherwise the system is running very smoothly and I am happy with it, but as it is running both hotter and louder than I would like it and as good a deal as these Phenom's are still don't want to break it.

[morphine]

Ah, I suppose one does learn something new every day. I had no idea that the Phenom IIs had such low thermal headroom. Which is doubly stupid because the average user never bothers cleaning out the heatsink/fan.

[Flying Fox]

I am not too sure if this is really the temp that these chips will go up in smoke. Throttling should happen before the smoking point, right? Call me confused.

[StuG]

The Phenom II temperatures are just the temperatures at which the warranty doesn't cover anymore. I have friends who have been running their Phenom II's up to 70C for over a year now, and not seen any issues. I believe that the chips can handle 70-80C fine, but that to ensure there are NO issues, AMD will not cover over 62C. Unsure why about this, but you may wanna check your general airflow of the computer. I have a SMF with a Phenom II 965 and Stock Heatsink that doesn't break 60C on 100% load on 8 hour Prime95 burn.

[just brew it!]

Just to add more confusion to this: AMD's maximum temperature spec is for what they call "Tcase" temperature. This is the temperature of the lid (heatspreader) of the CPU, where it contacts the heatsink. Since there's no thermal sensor located there, the reading is really only a guesstimate -- probably based on reading an on-die sensor and applying some fudge factors. So I'd take your temperature with a grain of salt.

[amphibem]

Thanks for everyone's help - can't say I fully understand the situation but sounds like I should be fine, currently folding on 4 cores at 61.5C.

As far as airflow goes, probably easiest to answer with some pics:



That is the back of the PC, showing an 120mm@5V sucking in air and a 120mm@12v on top of the heat sink pushing air out. The GPU has its own cooler.



View from the same place with case door on showing the exhaust vents. There is also another 120mm@12V on the front sucking in air over the hard drive.

The general principle is pull air in from the front and back, have it go through the various heaters, then exhaust out the side. However doing this has given me the idea of trying reversing the flow on the CPU and rear exhaust fans to see what that does.

[just brew it!]

The general principle is pull air in from the front and back, have it go through the various heaters, then exhaust out the side. However doing this has given me the idea of trying reversing the flow on the CPU and rear exhaust fans to see what that does.

It is certainly worth a try. You may find that the increased airflow you get by doing a "wind tunnel" type arrangement more than makes up for the fact that the air going to the CPU has already cooled some other components.

[Looking for Knowledge]

The general principle is pull air in from the front and back, have it go through the various heaters, then exhaust out the side. However doing this has given me the idea of trying reversing the flow on the CPU and rear exhaust fans to see what that does.

What case do you have?

I usually pull in from the front and push out through the top and back. Your config might have half of the hot air circulating around the heat sink at any given time.

[amphibem]

The general principle is pull air in from the front and back, have it go through the various heaters, then exhaust out the side. However doing this has given me the idea of trying reversing the flow on the CPU and rear exhaust fans to see what that does.

What case do you have?

I usually pull in from the front and push out through the top and back. Your config might have half of the hot air circulating around the heat sink at any given time.

Good point on the airflow - hence why I will try a different fan configuration. My case is a Thermaltake Wing RS 100, plenty big with lots of fans and side grills so should be able to cool anything with the right airflow.

[morphine]

Topic moved to OC/Cooling forum.

[amphibem]

So went with Looking for Knowledge's suggestion and got a reduction loaded temps, from 70+ down to 55deg. Still higher than I'd like but much more livable. Will now leave this box chugging away at GPU folding only, at the risk of getting my thread moved again found I get better PPD with just GPU or SMP (~3000) than both!

[spartus4]

I have a Phenom X6 1055T and it idles at 25.5C on all cores. When under load it doesn't go above 46C. I would think your's shouldn't be anymore than mine. Load up CoreTemp and check the Tjmax for your processor. Mine is 67C so I'm well below max. Also check your BIOS settings. Make sure your fan control is either set to Auto or Enabled. Another thing might be to see if there is a new BIOS for your Motherboard. Sometimes that can help. Best of luck to you.

[Crayon Shin Chan]

That's funny, my 1055T's Tj.Max according to CoreTemp64 is 99 Celsius.

[just brew it!]

That's funny, my 1055T's Tj.Max according to CoreTemp64 is 99 Celsius.

Tj.max is the maximum die temperature. The 62C figure quoted previously is maximum case (heatspreader) temperature. The die will always be hotter than the case.