APU in new PC is getting too hot

I got a new PC recently,  it came with an APU.  Its the A-10 6700 with the Radeon HD graphics built into it.  I noticed though that it would get really hot on every program that could read the temperatures. It had even different temperatures than what it was displaying on Speccy than it had on AMD Overdrive.  It was at about 130F and was randomly jumping up to around 160F and 180F during me using programs with graphics. 

The PC should be using my GT 640 instead of the one in the CPU and it is supposed to be according to my PC using my card.   The CPU also has a turbo core for 3.7 Ghz to 4.3 Ghz feature which I tried to figure out how to disable. I think that may be what was partly causing this.  I managed to find AMD Overdrive online to download and turned it off. 

I learned that there is about 3-4 different things on my PC I can adjust for the CPU  including AMD Overdrive, the BIOS, Catalyst and some power management settings.  It now runs about 125F-130F idle and seems to jump to at least 140F or even 150F during some graphic programs.

 I was wondering if anyone knew what is up with this because I'd think the PC should have been able to handle the APU temps without me having to do this.  Or if anyone knows a way to fix it.  Aside from trying to get better cooling or re-seat the the heat sink and putting new paste on which I cannot do right now.  Google seems to show other people also having issues with heat for it.

My warranty is technically voided since I swapped the PSU to use my GT 640 and put an extra fan in and wasn't expecting to have a issue with the temp.  Sometimes I think it might still be overclocking a bit because the multipliers and speed shifts when the temperature goes up too.

Hi.  I'm probably not the best person to answer this, but those temperatures really aren't too bad depending on the load you are putting on the system.

I would expect 45 to 55 degrees celcius during idle and up to 75 to 85 degrees celcius under full load with stock cooling.  This might be totally off the ball, but in my limited experience monitoring CPU temperatures it has not been unusual to see difficult to cool devices such as laptops get up around 80-85 degrees under full load.  I certainly wouldn't want it getting any higher than that, and I personally think this is a little too warm but it should still be within operating temperatures albeit probably very close to limits and probably risking thermal shutdown.

If you've added a video card you may be producing more heat inside the case and not evacuating it well enough.  I have to wonder if this is a custom built machine or a name brand machine?  If its custom built, then maybe something wasn't actually installed properly like the heatsink.  It certainly wouldn't hurt to remove the heatsink, examine the two contacting surfaces and see how the thermal compound spread out.  Maybe it isn't seated fully.  You'll just want to be ready to clean off the old compound and add new.  I guess this is possible with a name brand system too, except I would think that the more likely problem on a name brand system is just simply inadequate cooling / heatsinks.  They tend to skimp on things like this.

Well, anyways, as far as the CPU goes and the settings I don't have anything to tell you.  But, stock cooling should be fine for stock settings.  You should not have to disable any features to operate within manufacturer specifications.  If you are specifically overclocking / overvolting then yes you can have problems.  But, otherwise, the CPU should be able to do what it was intended to do without concern about heat with only stock cooling.

On a side note, it doesn't void your warranty to open the case and add hardware or replace components.  At the very least, you can remove the video card and install the original power supply.  If there is no physical damage due to your tinkering then there should be no problem replacing or warrantying a system.  

That is how it was when I got it everything was on normal power and what not but it was having alot of huge peaks randomly in the temp up to 160F or even 180F in some graphic intensive applications intill I switched it to power saver mode it stoped getting as high.  My grapic card only runs around 100F and consistaintly stays there increasing only about 10 degrees in games or less.  I never intentially tried to overclock the APU/CPU.  Its built with AMD Turbo core though to boost the clock.  Which I assume the multipliers are another name for a clock,  Everytime they increase automatically from x18.0 to x 23.0 or even higher is when the temps start to spike up, more so the higher the multiplier the higher the temp intill it goes down.  But it doesn't always go down when your running things that are putting load on it.  For example a game.

  It is a Asus M11BB brand pc with some modifications.  I swapped the 350w for my 430w psu so I could use my gt 640 which requires 350w minmum psu.  So I did not want to risk havingtoo much on the already 350w psu.  And added an extra fan.  I manged to get some thermal compound. I took off the heatsink, cleaned off the thermal compound on it and put on the new paste.  It didn't really make a difference in the temperature.

The fan on the heatsink blows away from it.  I was alsowondering if it might be better to try to flip it around so it blows on it.  However the screws on that fan won't seem to grip on any screw driver I've tried.  And its hard to unscrew it when the other screw hole in the fan is right above it that the screw driver would have to be thin enough to fit through that hole to get to the screw.

Here is example from Speccy of what happens in some spikes;(It is not the highest temp it has reached though since highest I saw was 190F once but only during running a game)

Closer to Idle:

 Core Speed Multiplier Bus Speed Temperature Threads
Core 0 1799.6 MHz x 18.0 100.0 MHz 135 °F APIC ID: 0
Core 1 1799.6 MHz x 18.0 100.0 MHz 135 °F APIC ID: 1
Core 2 1799.6 MHz x 18.0 100.0 MHz 135 °F APIC ID: 2
Core 3 1799.6 MHz x 18.0 100.0 MHz 134 °F APIC ID: 3

When it starts to spike:

 Core Speed Multiplier Bus Speed Temperature Threads
Core 0 3399.2 MHz x 34.0 100.0 MHz 152 °F APIC ID: 0
Core 1 3399.2 MHz x 34.0 100.0 MHz 152 °F APIC ID: 1
Core 2 3399.2 MHz x 34.0 100.0 MHz 151 °F APIC ID: 2
Core 3 3399.2 MHz x 34.0 100.0 MHz 151 °F APIC ID: 3

just to re-confirm what appleoddity said.

152F is nothing to worry about and is only 66C, perfectly normal for a pc under load.

my pc with watercooling hits those temps.

the a10 6700 max temp is 160F or 71C according to this :- http://www.cpu-world.com/CPUs/Bulldozer/AMD-A10-Series%20A10-6700%20-%20AD6700OKA44HL.html#specs

so i doubt the 190F you saw was the correct temp because the cpu would have shut itself down if it reaches it's tjmax of 160F.

if your really worried then i'd suggest swapping the manufacturers heat sink for an after market heatsink for better cooling performance.

another thing to take into consideration is some temp monitoring programs are known to be highly unreliable.

i'd try this one and see what your temps are. :- http://www.cpuid.com/softwares/hwmonitor.html

Wise words above, and to start with, I have to admit I saw this when posted, but the term APU threw a curve ball at me, I had never heard that term before, so I shied away from responding at that time, more so as a search of APU showed nothing that made sense to me, so I figured it was something I have never seen, well it is really, but after seeing the above, and more searching I finally found out what it was about, but I had never heard the term before this.

Anyway now I know what it is, I can expand on the above, but the main things to note are, software readings are not always that accurate, and the first thing I would suggest is checking the temp reading in the bios, see how that compares to the other readings, but due to the amount of variables affecting temperature it can be a case of more than 1 remedy is required, so to start with these are my opinions on temperatures (others will have theirs, but all would agree when things start to get dangerous);

I would expect to see the temperature at around 86°F (30 degrees C), up to 104°F (40 degrees C) in some cases.

When in full use I wouldn't be concerned if it reached the 140°F (60 degrees C) range, and that's under heavy load, although I would be starting to look to get it down if possible.

BUT, when it goes past 158°F (70 degrees C) I would start looking to better it straight away, I'm not saying it would cause damage, yet, but it would be the precursor in my eyes, and systems I touch I set the warnings in the bios to start once it hit's this temperature.

When it hits 176°F (80 degrees C), I would start worrying about physical damage, although the temp wouldn't immediately do this, prolonged running at such a level will certainly shorten it's life.

The last one I will note is 194°F (90 degrees C) and above, this is the point I would start to say turn the system off to protect it (and it's what I set the bios to do), and in some cases it "may" have already done some damage, but please don't think this is a definite outcome, there are far too many variables, but needless to say the higher the temp once your beyond this point, and the longer it's left, the more likely it becomes you could fry your CPU, and possibly other components as well, but again these are the things I personally go with, others my go a bit further, and in rare cases like the P4 series, they were renowned for running at very high temps, which can make a difference to these temps I use.

Now you sound like you have a good idea about what I cover next, but as this may help others reading this with a similar issue, I will mention them to be safe, to start with the screws you mention, when they get to that stage, you are at risk of stripping the heads, which means you would need tools to sort it out, and if you can get them out, do NOT put them back in, get new screws, as if you do it too much you will need tools to remove the screw(s), and it can be very difficult depending on your luck, so rule of thumb is if your screwdriver doesn't fit very well when you first use it, get a better fitting one, it will save you a lot of heartache down the road, more so if you like upgrading etc.

The Push\Pull query (air flow) is one of try it and see, again due to all the variables it could make a big difference, other times there won't be much in it, I remember experimenting many years back with this, and I found little difference, but I personally have the fan blowing ONTO the heatsink, but you will come across aftermarket heatsink's with 2 fans, one is blowing on (push), the other one on the opposite side of the heatsink away (pull), this compliments the other and helps in that product.

The biggest thing that can make a difference is the case, it's hard to picture it unless you have dealt with it, I had a mate with very high temp's, and they had a small case, which due to size was quite cluttered, I was lucky to have a spare case, so all I did was swap their parts into my other case, I did not touch or do anything more, and the temp's dropped by over 86°F (30 degrees C), so the design and airflow inside the case has a massive impact on this, although with newer cases these days that has become less of a problem. but I mention it incase.

Heatsink, the stock one that comes with it is always good enough to start with, otherwise it would never be there, BUT, you can get far better temp's by using an aftermarket option (that's my findings anyway), but if you are not happy with your temp's, then this is usually a great solution if you get a quality part, but be aware the better ones are massive, and may not fit in small(er) cases, so be aware of this, but if you intend to max the system out then it's always a good idea to get a replacement for the stock cooler, otherwise if the temp's are constantly high, then this is usually the first thing to try.

That said, I do mean if the temp's were constantly high, if you have normally been used to low\reasonable temp's, and they suddenly go up, reseating the heatsink with new thermal compound is usually the first thing to try, and when doing this it's essential to clean off the old BEFORE applying the new, and nearly always sorts it out, but only when it has changed from what it has been running at since you got it, or if 2nd hand is to be considered not knowing it's past history, being the cheaper thing to try first.

Fan's\Airflow, no prizes for guessing the impact this can make, but often people underestimate this, and don't put the fans in the best places, or the best way round, so in simple terms all fans should be pushing the air toward the rear exhaust fan, if your case has a front fan (or space for one), this should be pulling the air into the case, if you have top fans they should be pulling the air out of the case, side fans vary IMO, some are made to pull air out of the case, more so ones that have a "funnel" that goes next to the heatsink I believe, other times it's experimenting.

Placement, no surprises for saying keep the tower\unit away from heat sources, such as radiators\fires, even direct sunlight which some may not think of, if you have bottom fans, or simply a mesh in the "floor" of the case, then if on carpet raise it up, and put say some hardboard (or anything flat and puts a barrier between the carpet and the intake, hardwood floors don't need this) on the floor and sit the tower on that, it reduces the dust\fibres from the carpet being sucked up into the system.

Cleaning, open the case every so often, say yearly, and give it a spring clean, when using a hoover to suck out the dust, if you are going close to the board it would be wise to use an old pair of tights, and stretch and cut a piece to place over the end of the hose, that way worst case, and a jumper or something loose won't get sucked up, you can just turn the hoover off and put it back, if using compressed air be careful how close you get to the board for the same reason, and is best done outside unless you want a dust cloud in your place.

A general note, when going inside a case ALWAYS practice safety, shut the system down, remove the mains cable completely from the system (and battery if a laptop), even turn off the switch on the PSU if it has one (belt and braces as they say), and ALWAYS touch the side of the case before touching the component's, it reduces the risk of static killing the parts, it only takes a couple of steps on a carpet to generate a charge enough to do damage.

Then depending on what your doing, take notes of wire colours and positions, take pictures (most people have a digital camera these days, even if only on their mobile), this way worst case, and something comes loose, you should be able to put it back safely and easily.

I've probably missed a few points, and sorry for writing a novel lol, but hopefully this will give you things to look at and check, and see if anything makes a difference, if it doesn't, then the only solution is to improve the cooling, the heatsink being the place to start IMO, so hope this helps a bit anyway.

I wasn't aware of the maximum temp for the A10-6700 but I did some cross checking and it does appear to be 71.5 degrees celcius (160F).  This was surprisingly low in my mind as most electronic components are rated for 180F and beyond.

Anyways, with that in mind the temperatures do raise some question as to if it is cooling properly.  But, I would go back to my original statement that if the CPU is running at stock settings, and a stock fan and air is moving through the case properly it should be able to operate normally without harm.  I may be off base on this, but historically AMD processors have run hotter than Intel.  From the little research I did on this processor it does run hotter than average.  Its a pretty powerful CPU so with power comes heat.

In regards to the fan direction on the heatsink.  Again, I may just not pay close enough attention, but I don't think I've ever seen a stock heatsink fan NOT blow down over the heatsink.  So yours is backwards.  Usually the air blows down over and through the heatsink then out and over other nearby components to help cool them.  

Ultimately, if you are not having any trouble, and the system is not shutting down to thermal events I'd be inclined to not worry about it too much.  But, heat can reduce the life of components.  So, the cooler the better.  If it makes you feel more comfortable, you probably will need to get yourself an after market cooling fan and heatsink.  A nice one can probably be had for $30-$40.

Those final temperatures you listed appear to me to be absolutely normal…