4GHz FX 6300 Overclock Crashes Windows

[byalexandr]

Hey guys! So I have been doing some overclocking to my new FX 6300, using the AsRock 970M Pro3. So here's what's going on:

 

I overclocked to 4.0GHz and overclocked the boost clock to 4.2GHz. The first time I tried to boot into Windows my PC froze at the Windows logo and I had to reset. The second time I booted into Windows, and as soon as I started Prime95 it caused a BSoD with the error CLOCK_WATCHDOG_TIMEOUT, which means the clocks aren't in sync or whatever. So I changed the clock speed to 3.7GHz over the stock 3.5GHz and changed the boost clock back to 4.1GHz. Now here I am typing this. Couple questions:

 

Do I have a bad chip?

Is it the 4+1 VRM on my board? (although others have been able to go to 4.8GHz with my board)

Do I need to change the voltage?

 

I really want to be in the '4GHz cool kidz club' so please help me out. Also, the 4.0GHz OC is on the stock voltage and I am using the stock cooler, but my case is very well ventilated and the fan doesn't increase speed while I load the processor playing games for several hours. Any suggestions? Thanks in advance.

[Tedster]

That's pretty good going if you're okay with the stock cooler. What are your temps?

[byalexandr]

That's pretty good going if you're okay with the stock cooler. What are your temps?

I don't know exactly, but the fan never ramps up despite several hours of BF4, and I don't think the temps from CoreTemp are right because apparently my 6300 is running at 14C right now.

[Tedster]

I don't know exactly, but the fan never ramps up despite several hours of BF4, and I don't think the temps from CoreTemp are right because apparently my 6300 is running at 14C right now.

What's your ambient temp?

 

That 'reading' sounds way off as well.

[byalexandr]

What's your ambient temp?

 

That 'reading' sounds way off as well.

Ambient is 71F so that's 21.6C. I have a 92mm and 120mm fan venting the rear of the case, and a single 120mm up front. All of the case fans are running at full speed.

[byalexandr]

Anyone have experience with overclocking FX chips? I could really use some help...

[-rascal-]

Anyone have experience with overclocking FX chips? I could really use some help...

 

Ambient is 71F so that's 21.6C. I have a 92mm and 120mm fan venting the rear of the case, and a single 120mm up front. All of the case fans are running at full speed.

 

You are probably looking at the wrong temperature reading for the CPU...like many people here have. What program are you using to check your temperatures?

 

The ambient being 21.6*C, and the CPU running at 14*C is simply not possible on air or water cooling (unless you us something like Liquid Nitrogen, or Phase Change cooling). This goes directly AGAINST the laws of Physics, Electrical Properties, and Thermodynamics (2nd general law) that governs our world...

 

http://linustechtips.com/main/topic/370500-stable-overclock-for-the-amd-fx-8350/

It looks like you are reading the temperature incorrectly there (like many people here with "Bulldozer" / "Pildriver" Processors). Heh, you are the...3rd of 4th in the last week or so.

 

http://linustechtips.com/main/topic/369522-cpu-temps-not-showing-properly/

 

In HWMoniotor, you are looking at the wrong thing.

The Package temperature is not accurate during idle, but should be relatively accurate when the CPU is under load.

If you want to see a more accurate reading, scroll up a bit in the HWMonitor menu.

This is another fourm member's CPU -- he also was making the mistake by looking at the "Package" temperature.

Under "Temperature" there is a reading actually labelled as "CPU." This is basically bang on with what the BIOS shows. If you look, my "Package" temperature is also very low...my room temperature at the time was around ~15*C).

For the heck of it...

Regarding to AMD OverDrive, depending on where in the software you are reading the temperature, it will show the margin you have left before reaching the thermal limit. So as your CPU heats up, the temperature will actually decrease...as it is showing the headroom you have left.

Comparing my "Temperatures" > "CPU" temperature reading from HWMonitor, with my own "FAN0" temperature reading in AMD OverDrive (installed it to post on this topic), they are identical. If your "FAN0" is your CPU_fan, then that it is likely that IS your CPU temperature (39*C)

 

 

http://linustechtips.com/main/topic/362783-cpu-problems/

 

Bingo!

Thank you for the video. (BTW, screen capture -- use the "PrtScn" key on your keyboard)

 

HWMonitor's "Temperature" --> "Package" gives incorrect readings when the system idle.

Rum a stress test, like AIDA64 or Prime95, then it will give more accurate readings.

 

If you go to 0:55 in your video, with Gigabyte motherboards, "Temperatures" --> "TMPIN0" (or TMPIN1) is your CPU temperature.

Your FX-4130 is sitting at ~31*C.

 

 

The other two is motherboard temperature and CPU/NB temperature (I don't recall which is which exactly for Gigabyte boards).

..

..

 

 

http://linustechtips.com/main/topic/363153-cpu-temps-100c/

 

Depending on your motherboard / CPU, HWMonitor may not explicitly have a temperature reading labelled as "CPU", "motherboard", or something like that.

It can be named as "TMPIN1" or "TMPIN2."

..

..

..

 

 

 

If the system crashes, and BSOD's, then your overclock it not stable -- especially with the CLOCK_WATCHDOG_TIMEOUT error.

You most likely need to increase the Core Voltage up a small amount if you want to have 4.0GHz stable.

 

I don't know about your temperatures for the time being...as you need to get an accurate reading.

 

Other then that:

• The 4+1 power will be a limitation on how far you can overclock, and how much MORE voltage you will need to keep the higher overclocks stable
• The 4+1 will not provide as stable, constant voltage versus something like 6+2 or 8+2
• The heatsink over the VRMs are quite small, and because they are so close to the CPU socket, it will transfer some of the heat over to the Processor; hence, increasing the temperatures a bit
• Due to the small heatsink passive cooling, the VRMs will run hotter

 

We don't know whether or not if it is a bad chip at this point, because the programmed stock voltage varies between CPUs.

[byalexandr]

You are probably looking at the wrong temperature reading for the CPU...like many people here have. What program are you using to check your temperatures?

 

The ambient being 21.6*C, and the CPU running at 14*C is simply not possible on air or water cooling (unless you us something like Liquid Nitrogen, or Phase Change cooling). This goes directly AGAINST the laws of Physics, Electrical Properties, and Thermodynamics (2nd general law) that governs our world...

 

http://linustechtips.com/main/topic/370500-stable-overclock-for-the-amd-fx-8350/

 

 

If the system crashes, and BSOD's, then your overclock it not stable -- especially with the CLOCK_WATCHDOG_TIMEOUT error.

You most likely need to increase the Core Voltage up a small amount if you want to have 4.0GHz stable.

 

I don't know about your temperatures for the time being...as you need to get an accurate reading.

 

Other then that:

• The 4+1 power will be a limitation on how far you can overclock, and how much MORE voltage you will need to keep the higher overclocks stable
• The 4+1 will not provide as stable, constant voltage versus something like 6+2 or 8+2
• The heatsink over the VRMs are quite small, and because they are so close to the CPU socket, it will transfer some of the heat over to the Processor; hence, increasing the temperatures a bit
• Due to the small heatsink passive cooling, the VRMs will run hotter

 

We don't know whether or not if it is a bad chip at this point, because the programmed stock voltage varies between CPUs.

I know the temps readings are wrong, as I know the CPU can't go below ambient without a substance like liquid nitrogen or freon cooling, which is not the case as I have a stock cooler. I will see if I can increase the voltage a bit and get a stable OC.

[byalexandr]

Wow... I screwed up my PC HARD. So here's what happened:

 

OC'd to 4.0GHz (3.5 stock)

OC'd boost clock to 4.2GHz (4.1 stock)

Overvolted one increment. I can't remember the exact voltage but I think it was 1.1625 or something, but nonetheless it was one increment above stock.

 

Restarted, BIOS crashed, restarted, BIOS passed, Windows loading, started flashing the screen, screen momentarily flashes with 'Preparing automatic repair', screen goes black again, continues flashing from Windows to black, restarts, stuck at BIOS, continues to..... You get the point. It was awful. I shut off the PC and cleared the CMOS, and then Windows repaired itself, and here I am typing this.

 

Well, that scared the crap out of me. I think I will stay at 3.7GHz, as I'm not the best overclocker (in fact, I'm far from the best) and it seemed to be running normal at that speed with no overvoltage.