AMD FX-8350 overclocking

nucLeaR_gogo

Reputable
Jun 4, 2015
21
0
4,510
Hello. I have bought my AMD FX-8350 two months ago, even before I've got it I was so excited about overclocking the chip and I've already watched a ton of tutorials, like JayzTwoCents' tutorial. I'm using an Arctic Freezer 7 Pro rev. 2 cooler. I really want to overclock it and I've tried it already, I've upped the frequency to 4.2GHz with a little more voltage and it looks stable when I stress test it, but it crashes my computer after 15 minutes of Prime95 testing. I dont get an "Overclocking failed" screen message or anything it literally stops the display output and everything else (fans, hard drives, etc.) stays on, but it just won't boot until I restart it manually. Then it boots up and everything is like normal, but the thing is I get so scared when it crashes like that and recently I've read some threads about people who overclocked their CPUs and after a few months they die (the CPUs not the people lol). I want to overclock it even if it shortens it's life to a few years. My temps are 61C when stress testing on stock clock speed and 65C with the overclock. I'm going to get an Arctic Liquid Freezer 240 AIO water cooler and when I get it I really want to overclock my chip to say, 4.6 or 4.8GHz maybe? My question is, what is the likelyhood of my CPU dying on me or my MoBo failing? I've already heard stuff like "There is no 100% safe overclock" or "There is always risk when overclocking", but my question is what is the chance of something like that happening, because I don't want to overclock it and after a month to not be able to boot my PC at all. BTW my MoBo is MSI 970 gaming. Thanks for the help!

CPU - AMD FX-8350 + Arctic Freezer 7 Pro rev. 2
RAM - 8GB DDR3 @ 1600MHz
MoBo - MSI 970 gaming
GPU - R7 260X 2GB @ 1050MHz
HDDs - 1TB WD Caviar Blue + 250GB Seagate Barracuda
PSU - Thermaltake Berlin 630W 80PLUS
 
I had an 8320 that i could get to 4.0 without changing voltage and in theory you should be able to get it to the turbo clock without changing the voltage either. Depending on how much you up the voltage, it could last a couple months to 10 years if you are lucky, if you up the voltage, its more likely to last at least 3. As long as you don't go crazy with the voltage and check out overclocking guides on overclock.net and other places as well and you follow their advice, you should be fine, just know there is still risk of having a bad apple and the chip might not be able to handle it, but that is the risk you take.
 


Well I have read that people overclock their chips to 4.4GHz with the stock voltage. What's wrong with my chip then ?
 


Here is a screenshot I made after 25 minutes of P95 stress testing on my FX-8350. I think the voltages are OK, but you are the expert.
XSFlb42.jpg


 
First of all, listen to blackbird. You PSU is a poor choice for overclocking. Second, if you're just gamine, no point of overclocking your CPU when you use a 260x. The bottleneck will always be at the 260x, OC'ing cpu won't make a difference.

With that said, if you want to overclock just for fun, then I understand. For monitoring temperatures, ditch HWmonitor and use AMDOverDrive, and possibility Speedfan. I mean, it is very awful when one of your program reads 51c and the other 60c. AMDOverDrive, gives thermal margin, which is the most accurate for AMD under load.

If the system crashes at any point you just finished overclocking, whether is testing, booting, gaming, or anything, then it is a sign of instability.

Btw, are you running Small FFTs or Blend testing?

 
Actually 52 is the socket temp and 60 is the core temp. And yes I know the bottleneck is going to be the R7 260X even if I overclock the CPU. I'm going to be getting a R9 390 soon. I am running small FFTs. I have also looked in AMD Overdrive. While HW Monitor and Core Temp report the same 58C while stress testing, AMD Overdrive says the thermal margin is 12C. Wasn't 70C the max. socket temperature? I thought AMD Overdrive measured the core temps.
 
I just don't like HWM and I don't know the other one you're using, the HWM always give that Max temp of about 200c. AMDOverDrive is the most accurate at load, just don't let it hit 0c and you'll back fine. If you're able to run 95 for 30-60 mins without any cores failing then you're fine.

Here is an article to AOD, little dated but principles are same,

http://www.tomshardware.com/faq/id-2122665/understanding-temperature-amd-cpus-apus.html

62c max core, 70 max socket.
 


The other program is MSI Command Center, I think it shows the most accurate socket temps. But I still dont get it. Usually the socket temp must be higher that the core temp, right?