Some help with first time system build - FX-8320 - weird temperatures

[Archonius]

Hi all! Recently decided to start building my own computer and decided to go with pcpartpicker's recommendation for a cheap but decent PC. Been having some problems and have been pulling my hair out so hopefully you can help Anyways, here are the parts:

AMD FX-8320, COOLER MASTER Hyper TX3 cooler
Biostar A960D+ (mATX)
Crucial Ballistix Sport 2x8GB
Cooler Master 620W PSU
Powercolor HD7770 Video Card
1GB Seagate HDD, 7200 RPM
Rosewill mini ATX case (92nm inflow fan in front, 80mm outflow fan in back)

So here is my problem. After playing a game like Battlefield 3, or Skyrim, for a few minutes, I notice the game starts stuttering for a few seconds, then goes back to being smooth. Then every 20 seconds or so, it will stutter for 10 seconds, then go back to smooth for 20 seconds, keep repeating.

I investigated using Prime 95, and having CPU-ID and HWMonitor open while I ran Prime 95 stress test. HWMonitor shows these in the temperatures section:
THRM
TMPIN0
TMPIN1
TMPIN2

In a separate section is
AMD Package temperature
Radeon HD7700 series temperature

Anyways. Now lets get to the stress testing. First, here are the temperatures on idle, about five minutes after booting into Windows and not doing anything, with room temperature about 75F (~24C):
THRM ~ 16-18C
TMPIN0 ~ 17-18C
TMPIN1 ~ 42-45C
TMPIN2 ~ 31C
Package temperature ~ 16-18C
HD7700 temperature ~ 34-36C

Now I open Prime 95 and start the program, here are the temperatures
THRM ~ 26-40C
TMPIN0 ~ 26-40C
TMPIN1 ~ 60-65C
TMPIN2 ~ 31C
Package temperature ~ 26-40C
HD7700 temperature ~ 34-36C

Here is what happens. I didn't overclock, and CPU-ID is showing a processor frequency of 3500MHZ which is standard for my CPU.

When I run Prime95, the THRM/TMPIN0/Package all climb relatively quickly to 40C, but TMPIN1 temperature gets to 60C, then slowly climbs to 65C. The second TMPIN1 hits 65C, CPU-ID shows that my processor frequency drops to 1400MHZ. When this happens, THRM/TMPIN0/Package quickly drop to 26C, while TMPIN1 slowly goes down to 60C. The second it hits 60C, CPU-ID shows my processor frequency going back up to 3500MHZ, while THRM/TMPIN0/Package quickly hop back to 40C, and TMPIN1 goes slowly up to 65C.

The frequency at which this "throttling" occurs coincides perfectly with my experience when playing BF3/Skyrim.

So here is what I think is going on. TMPIN1 is my actual CPU temperature, not "package". It has some throttling system set up in BIOS which means at 65C, it will throttle down to save itself. So either my cooling system sucks, or something else is wrong. I really need a way to figure out if this is true, that TMPIN1 is my actual CPU temperature, because even though it behaves like it when the computer is working, the labels on temperature monitoring programs, and hell even BIOS itself make it look like the THRM/TMPIN0/Package are the CPU temperature.

I would appreciate any help or advice you can provide. Thank you

 

[Archonius]

OK so I just tried putting the stock AMD cooler back on, and got the exact same results but WAY louder. So I don't the Cooler Master heatsink is the problem.

 

[cageymaru]

I have a FX-8120 and an Asus Sabertooth 990FX version 1.0 mobo. When using my hardware monitoring software that came from Asus, I notice that my V-Core-1 and V-Core-2 temperatures would go crazy. Then soon my cpu would hit 50c - 55c and I knew that as soon as it crept to 60c that the computer would restart.

Basically my vrm were overheating way before my cpu could break a sweat even though Asus has nice heatsinks on them. The heat coming off them was causing my cpu to thermal and cause stuttering then crashing.

My solution was to remove the stock cpu cooler's fan and mount it near the vrm. Instead of them boiling my cpu, now they idle at 34c.overclocked.

Here is a post from someone else showcasing a solution. You will see what I mean.

 

[ob1]

You could also try removing the side panel (if possible, haven't looked at your case), and put a house fan flowing into the case as a quick and dirty test. If the problems go away, you know have have a heater/airflow issue.

 

[Archonius]

Thank you cageymaru and ob1. An update: I turned off the temperature control thing in BIOS so it no longer throttles. Played games for an hour or so and temp of TMPIN1 got up to 85! Was pretty scared. Saw your suggestions about airflow - what I did was flip the rear case fan around. Before it was sucking air from inside and blowing it out, I just flipped it around so now it takes air from outside and blows it in.

What TMPIN1 is I think, mosfet temperature. This is a 3+1 board. With the fan turned around, it blows cool outside air directly over them and into the cpu heatsink, which is already being 'underused" since it is quite cool. the cpu heatsink fan is below, and it blows air up, and then the PSU fan bottom side takes that air and blows it out of the system.

temperature under load now - 68C. still not the best but 68C is much more comforting than 85C+!!! i dont care if this mosfet melts really. it is a $40 motherboard and i am just going to buy a better, decent one with 4+1 at least.