• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

FX Thermal limit

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

NiteSmoker17

Member
Joined
Jan 20, 2002
i was surfing the web and ran across this on svc's site
http://www.svc.com/am3-x1.html

now its for a thermal pad but what sparked my interest was the caption

AMD’s specification of 61°C-70°C is for maximum allowed case temperature (a thermocouple measurement taken on top of the CPU lid); however, one will be measuring CPU core/junction temperature with a utility such as SpeedFan, etc. As there is significant thermal resistance between the CPU lid and core/junction (due to TIM1, etc), the core temperature will be much higher than the case temperature. Therefore, it is possible to reach the recommended ~80-85°C average core temperature, while not exceeding the AMD max case temp of 61-70°C.

does that mean that we are looking at 85°C for max die temp for the FX chips? I'm asking this because i personally heard of a max temp for the FX chips.
 
Using the free edition of HWMonitor we normally suggest the package temp be kept to about 60c and the socket/cpu temp to about 70c. So far that has held up pretty good, otherwise early boards would throttle at or often below the socket/cpu temp. Those were the boards without APM disable setting in bios.
RGone...
 
I cant say for sure that the FX CPUs enjoy going that hot, but I have personally pushed one of my 8320s to 80C+ core temps for and extended period of time without a crash. Its definitely alot tougher to keep a good overclock stable once the temps start to get past the 65C range though.
 
I cant say for sure that the FX CPUs enjoy going that hot, but I have personally pushed one of my 8320s to 80C+ core temps for and extended period of time without a crash. Its definitely alot tougher to keep a good overclock stable once the temps start to get past the 65C range though.

Sure is tough. Takes a good VRM circuit for sure. The high heat increases the internal resistance and it is possible that the one's get lost in the fray.
RGone...
 
Well i wasn't sure if i was reading that info correct or not. its harder when AMD doesn't give us information to know the limits of the chips and i know i have had mine to 72°c for a short time doing prime95 but not for extended periods. just cause the woman would kill me and then i wouldn't have a pc. but seeing that info i wasn't sure how much weight was towards the FX as opposed to the phenoms and previous gen chips.
 
these fx cpus will run at very high temps.
I have seen my "package temps in the upper 80's, it's unstabe but it will take it all day long.
I had a pump fail when i was out of the house all day and that thing was just glowing hot when i got in.
I still use that cpu.
 
Okay this topic has just really shocked me to have read so far....

Hasn't the thermal limit of the "CPU" reading for the AMD 125w chips been claimed at 61c ever since the PII X4 lineup was released?! If I can get the chip past 60c all the way up to 70=80c(!!??), that would mean that I could be running something like 5200mhz on air cooling alone. Given that I've had to kill myself trying to get up to 5ghz without overheating on air and haven't seen many others go even that far, I can't believe anyone is really pushing to those limits mentioned above :shock:

Would anyone really feel like it's safe to push the temp into the 70's for a few minute of benchmarking with enough confidence to suggest it be something I let happen?
 
I don't think that there are more then a couple pushing those upper limits here, if there is they have kept quiet about it. CD is one of the few that openly talk about it, for me the highest I have pushed it was 65*C core and 75*C socket (or there abouts) When pushing 5GHZ when the 8350 first came out ..... that was 3 or 4 hours stable on prime. Please keep in mind I am running a very capable custom Water Loop for cooling my CPU.

I have since turned down my OC to stay in the recommended limits we have been suggesting to people here in these forums of 60*C Core and 70*C Socket. The biggest question or unknown about running these chips this hot is what harm, long term is being done to the CPU, only TIME will tell.
 
Okay this topic has just really shocked me to have read so far....

Hasn't the thermal limit of the "CPU" reading for the AMD 125w chips been claimed at 61c ever since the PII X4 lineup was released?! If I can get the chip past 60c all the way up to 70=80c(!!??), that would mean that I could be running something like 5200mhz on air cooling alone. Given that I've had to kill myself trying to get up to 5ghz without overheating on air and haven't seen many others go even that far, I can't believe anyone is really pushing to those limits mentioned above :shock:

Would anyone really feel like it's safe to push the temp into the 70's for a few minute of benchmarking with enough confidence to suggest it be something I let happen?
OC I have had mine over 70c on the package/cores running benches, I believe CineBench was the one, albeit, a short period of time. I am not trying to blow up my CPU but I'm not all that careful with it either. Would I recommend it to someone, sure if there is an understanding that it's no-mans land and we really don't know what the effects may be.
 
Use any Temp you feel comfortable with...Your parts.

This is as of early April 2013 not sometime in the distant past.

Motherboard >> ASUS M5A97 LE R2.0 ATX AMD Motherboard with UEFI BIOS
CPU >> Amd FX-8120
CPU Cooler >> Cooler Master 412S

Question:
Can anyone please tell my how to disable CPU throttling when its reaching 60C~?
Ive disables pretty much every power setting in BIOS and got good air flow inside the case with 4 fans. Everytime it reaches 60C~ the CPU downclock itself, really annoying. The "core temps" are at that point just under 40C I think. So if its the "core temp" I should look at, why does this happen?

Answer:
Says it is able to run 8 core/140 CPU's however it appears to be a poor marketing issue why?

Well the $89.00 boards from asus are good quality entry level boards in my experience meaning you get what you pay for.
1. > The majority of entry level motherboards have weak power phase design 4+2 Asus uses quality VRM components but in this case I would not recommend a 4+2 phase non-heatsinked board to run a FX 8 corel
2. > We use the term weak when compared with the PRO or EVO edition of the M5A97 that has a 6+2 phase heatsinked design or the even higher performing line that uses 8+2 phases such as the Crosshair or Sabertooth 990FX.

Now to answer your first question it appears your bios does not have a feature to turn on or off over current protection, higher end boards under the digi+ vrm/power control menu have a feature that can be enabled or disabled called VRM over temperature protection the features you have likely disabled are really only to save power/reduce temp during idle stages such as C & Q, C1E, EPU etc... or allow sleep functions like C6 core state

Your system is throttling (dropping 12v voltage/multiplier) because the VRM components are over heating, power is being cut to the processor to ensure no damage to the PWM system or the CPU
The FX 8 core processors draw immense power (http://www.bit-tech.net/hardware/cpus/2012/07/27/amd-fx-8120-review/7)

Suggestion: With the 8120 you need OCP or you will eventually fry components....drop down to a 6 core or add passive cooling to the all the mosfets 4 per phase using MOS- C1's or the like. If using the stock cpu cooler airflow is being provided to the VRM components, however if using a tower cooler such as the Cooler Master 412S listed in your sig airflow is blocked requiring you to use a extra fan to cool this area.

Anyone who comes into the OCF AMD CPU Forum section and either gets help and stays and helps day in and day out or who comes to help and never asked for or sought help; it is these people that day in and day out that must deal with users and the cheap stuff. It is those people that deal with exactly the posting user above. In real life the questioner has little chance of overall success with his choice of parts and pieces when run as HE "thought" he could utilize his parts and pieces.

So when Bulldozer began to surface back in the fall of 2011, there were post after post of users asking the same question as above. Many boards were throttling due to VRM overheat and even UNTIL today there are cheap boards that throttle due VRM load that can be related to CPU socket temp. By the time the CPU Temp gets there the VRM has throttled the cpu. Did the cpu socket temp do the actual throttling? No, but the temp caused the need for higher Vcore and then you are in the more temp, more voltage more temp circle of trying to balance an overclock.

Previous to the FX release most of us that stay day after day, were suggesting 55c Core temp and 62c socket temp. As time has gone on and the good boards features have gotten more diverse, we have upped the suggested Max socket temp to ~70c and the suggested core/package temp to ~60c as viewed in HWMonitor free version.

And now we get questions about this BS at this link.
http://www.svc.com/am3-x1.html
First they say this >> AMD Phenom & Athlon II/AM2/AM3 Sockets Only << that is not FX at all.

Then they say >>*Please note AMD FX Processors and Indigo Xtreme AMD AM3-X1 re-flow issue << still not AM3+ stuff. AM3+ stuff is FX processor stuff.

Then they say >> AMD’s specification of 61°C-70°C is for maximum allowed case temperature (a thermocouple measurement taken on top of the CPU lid) << which AMD has said in open forums is NOT the truth.

And then they finish with this >> In the mean time we are declaring Indigo Xtreme AMD AM3-X1 -AMD AM3/FM1 as possibly incompatible with AMD FX Bulldozer series processors until we can confirm/disaffirm this. << Good idea since nothing about the product has said it applies to the AM3+ series of cpus which are the FX processor series of cpus.

The entire thing at SVC was BS and useless except you intended to buy Indigo Extreme TIM for FX when they said now don't do so. Did not plan on buying the stuff anyway.

So the individual that wants to run 85c core temps...have at it. Go for it. Heck run any temp one wishes to run. After all you are the one buying the parts and would have to stand for any failures. No one in the forum will stand for your failures. Don't let a little caution and or suggestion put you off from some air-cooled clock you think you can reach. Go for it. Get 'R Done.

For the rest it is still suggested by those that help everyday that you use some logic that has been developed over the last year and half of dealing with the newer AM3+ cpus, and that ~60c Core/Package temp and ~70c Cpu/Socket temps are still considered pretty safe and sane as max temps. Such for many reasons that are not just related to cpu life but having cheap boards that can throttle at low Socket temps and that to be able to push the high current thru the VRM you must DISable the safety features are some of the other safe and sane reasons for using the ~60c and 70c as Max temps for Core/Package and Socket/Cpu temps.
RGone...
 
It has been said already but needs to be emphasized that those high temps quoted from the AMD statement are for stock frequencies and voltages. The stability threshold for a significantly overclocked CPU will be considerably less and I mean thoroughly stress tested when I say stable..
 
Last edited:
Back