People who own an FX 8-core:
How do you stop it from thermal throttling? It drops down to 1600mhz and 0.9v after it reaches 65c. I looked for anything related to this, disabled turbo, cnq, a couple other things.
What can be done?
Anywhere I look people just say get a better cooler but I already have a corsair h105.
Thats a decent vcore for a 4.5ghz overclock. That AIO loop would have no problem cooling it.
The chip shouldn't be throttling at all, which would lead me to believe that its the VRM on your mobo which is having an issue.
Btw, my motherboard is this
It doesn't 'officialy' support fx 8350 but it supports bulldozer (8150), and works with my 8350 fairly fine. I guess overclocking might be a problem?
Are you sure you're looking at something that reports temps correctly? Speccy shows 30c for me while ai suite/hwmonitor show 60c, for example.
Anyhow.. it still seems high idk why it gets this high. I have a decent aio cooler and case full of fans.
you should be good when it comes to vrm.
have you updated your bios lately (2012 was the was when last version released)
im going to /r/ a pic of the guts
oh and reseat your cpu and you may just see the problem go away (or just see a problem.)
I had to update it for the fx processor to work. it's the latest version.
Best pic I got I think, there's 2 more fans on the right for intake, and 1 on the bottom
I applied thermal paste again on the cpu a couple days ago. Should I take it out and in?
don't tell me you use a shitty asus with 3+1 phase.
because the vrm will throttle from overloading.
if you do have a shitty mobo, i suggest keep it stock clock and undervolt. at least it will run efficiently and cool.
Posted my mobo here >>51415745
Just tried testing after changing this option to 100%>>51415941
Still fucking throttling, this time at 67c
Look at max temps, does anything indicate this is the vrms?
I've repasted it very recently
It would overheat instantly if it was defective, and I had bad contact before and it instantly overheated
Could it really be half-bad contact or some shit?
67c is really fucking high for an 8350, or any FX chip really.
These readings are wrong.
67 nevertheless is hot though
>tfw you have the only FX processor with a functioning temp chip
So I reseated it, and reset bios to default.
These temps are with prime95 running, but fans are at 100% (they were dynamic before but set to 100% at 63c)
For some reason voltage is 1.25 or so
Could ram affect this? I've read that fx is funky with 1866mhz or even 1600 ram, and maybe my old mobo too.
Wrong reading again
Check your motherboard software for accurate reading.
>mfw im stable with a shitty 3+1 mobo at 1.36v
>mfw running like this for like a year now
4.5GHz, I turned on that speedstep or whatever AMD's one is called a while ago and I haven't bothered to turn it off.
All your temp is ok. I would avoid prime95 for FX as my previous fx6300 still downlock while stress test.
You can try running extensive game with unlock fps and check your cpu still downclock mid game. Try use osd from afterburner.
>it reaches 65c
There is your fucking problem right there. Buy something other than the stock cooler.
Desktop processors should never be throttling, and its not AMD's fault, its yours.
My 8320 is clocked to 4.5ghz on a corsair h110i GT(fuck this cooler btw). The absolute max temps is 59c for me. When using Aida64 system stability test, it can range from 53 to 57 depending on the temperature inside the room. When gaming it's usually in the 40s, newer games tend to bring it up to low 50s. Getting the best tuned overclock is not easy. Took me 2 years to get this shit to 4.5. Getting it to like 4.2 was no big deal and setting the perfect voltage wasn't either. If everything is set perfectly, your temps should be like low 30s on idle and 50s on load. Make sure you seat the cooler correctly with the right amount of thermal paste. Having good airflow makes a big difference as well. That'd pretty much it.. Don't use prime95 btw, terrible for amd chips. I recommend Aida64 or even Valley benchmark with everything turned up( represents what to expect from a very intensive game).
All in all, just reseat it properly or change the coolers fans. Can't wait to get rid of this chip. Time to finally get an Intel.
The 2 fan connectors built into the block don't even work. As I was testing them it hit the back of my mobo and shorted it out. When I say they font work, I mean in the sense that fan controls are buggy and fucked up.
The cooler along with the 2 fans(phanteks fans) vibrate my case randomly and the sound can get very annoying. Font need to mention stock fans suck when it comes to noise and cooling.
When it comes to actual cooling, it's not that bad at all. It's just for the price I never expected these problems at all,might as well get something else for the money. Now I must buy rubber gaskets as well as rubber screws for the fans and some sound dampening foam from Silverstone. The foams not necessary, I just need it for something else, so using it on my pc is just a bonus. Here's a pic of my most impractical build ever lol.
Used to be industrial double sided tape but after a few months, it started to lean to one side. Now I just have both stacked on top of each other and secured with the same tape. So, double sided tape on the bottom of the case>aio>more tape>top aio.
Yea, that's what I was thinking as well. I'd get a warning on my asus motherboard that the temps have reached 65c. Then at around 73c or so it would just downclock to something like 1.4hz until I restarted the pc.
This is my 8350 with a D15 on nothing but air cooling.
I'm not op. Just sharing my experiences with the stock cooler. Idk if it's the same for other boards but I would get the notification popup from asus suite 2 at 65c. It wasn't until 73c or higher when it would finally downclock. It's a safety measure, not a malfunction
Shit that AMD owner's need to go through....
Fuck that.. Let me be happy with my Intel and Nvidia components, I dont have to worry about anything getting hot or anything, in fact I dont hear anything in my computer at all
I am glad 0.5 GB vram is actually slower, this prevents the GPU from overheating itself by the high RAM clockspeeds
Nivea did such a great job engineering this GPU, nivea really is the way it's meant to be played
Then you know you're wrong.
Also one of my friends uses haswell, during a stress test his cpu got to 95 degrees Celsius.
During battlefield 4 it sits around 80 degrees.
His gtx 960 definitely runs cooler than my hd 7950 though.
You're a fucking shill,or you're an idiot.
I cant believe noone has mentioned this yet.
YOU MOUNTED THE COOLER WRONG OR YOU PUT TOO MUCH THERMAL PASTE.
This causes overheating+throttling. A cooler like that should be able to handle this CPU no probs.
I ran a 8350 on a 212 EVO and it never went above 58 on 4.5 GHz.
Haswell runs stupid hot during the normal stress testing go tos of past like Prime 95. No one that has any sense uses stress tests that run that much higher than BF4.
I used the Intel burn test.
He has a haswell xeon, got it instead he an i5 since he doesn't need the igpu.
In comparison, my i5 2500 peaked at 53 degrees in the same stress test, but not on stock cooler.
I have a cheap $25 heatsink, and it never goes past 55 while rendering video.
I think it has something to do with the addition of AVX2, it runs it abnormally hot. Even though IBT already does that in the first place. Also why newer versions of Prime 95 aren't recommended either.
You must have messed up somewhere. My cpu doesn't go over 60 with the default cooler while gaming for long amounts of time. Either you live inside a volcano and need to open a window or reapply thermal paste.
the only valid way to check the core temps is with AMD OverDrive
install it, and if you're never hitting the "thermal margin" in there your processor and cooler are working perfectly, your problem is the mobo temp around the dice or the vrm
maybe the solution would be to install a C type cooler like the noctua nh-c14 to cool everything
but of course, the smart solution is to get a fucking Intel chip+mobo for gods fucking sake