Question Help with OC'ing FX 6300

Jan 28, 2020
91
8
45
2
Hello,
My system specs are: FX 6300 w/ Stock Cooler, M5A78L-M/USB3, 2x4GB 1600Mhz RAM and GTX 750 GPU
The temperatures are around 62C and 64C under 30 mins of stress test and an idle temperature of 21.4C and 37C on package and motherboard CPU sensor respectively.
VCore seems to jump upto 1.42V at 4.12GHz turbo. So I decided to try undervolting the CPU, but the BIOS only have option to increase the voltage over stock voltage for the CPU(i.e 1.3v) but not undervolt.
Should I just leave the voltage on auto and disable Turbo Boost or should I overclock the CPU with voltages under 1.375 or so?

What are my options on this? Any suggestions would be really grateful.
 
Hello,
My system specs are: FX 6300 w/ Stock Cooler, M5A78L-M/USB3, 2x4GB 1600Mhz RAM and GTX 750 GPU
The temperatures are around 62C and 64C under 30 mins of stress test and an idle temperature of 21.4C and 37C on package and motherboard CPU sensor respectively.
VCore seems to jump upto 1.42V at 4.12GHz turbo. So I decided to try undervolting the CPU, but the BIOS only have option to increase the voltage over stock voltage for the CPU(i.e 1.3v) but not undervolt.
Should I just leave the voltage on auto and disable Turbo Boost or should I overclock the CPU with voltages under 1.375 or so?

What are my options on this? Any suggestions would be really grateful.
You are limited by your stock CPU cooler and VRM on the motherboard to some extent. There are cheap but much better CPU coolers out there , https://www.lc-power.com/en/product/cpu-cooler/lc-cc-120/ for instance.
 
Jan 28, 2020
91
8
45
2
You are limited by your stock CPU cooler and VRM on the motherboard to some extent. There are cheap but much better CPU coolers out there , https://www.lc-power.com/en/product/cpu-cooler/lc-cc-120/ for instance.
Yes I know I am limited with the stock CPU cooler, but I do have a 80mm fan blowing towards the VRM. So you mean to say that there's no option that I can bring down the voltage and have a stable OC under 4.0GHz which in-turn reduces the temps in the system?
 
Yes I know I am limited with the stock CPU cooler, but I do have a 80mm fan blowing towards the VRM. So you mean to say that there's no option that I can bring down the voltage and have a stable OC under 4.0GHz which in-turn reduces the temps in the system?
You can't have a cake and eat it, higher frequency simply needs higher voltage to keep it stable.
 

zx128k

Upstanding
Nov 23, 2019
474
88
290
6
Are you thermal throttling? Stock coolers are mostly about getting the job done and not about keeping temps to a minimum. Also if you have been running the system for years, have you cleaned out the dust from the heatsink and fan?

The max Boost Clock for the FX 6300 is up to 3.8GHz AMD. http://www.cpu-world.com/CPUs/Bulldozer/AMD-FX-Series FX-6300.html shows,

Boosted P states [1]#1: 4100 MHz, 1.425V
#2: 3800 MHz, 1.4125V

So your 4.1GHz @ 1.425 volts is normal.
Maximum operating temperature ?70.5°C

You can check your thermal margin as well if you like.
The Tcase Temperature of FX 6300 processor is 70ºC. source email from amd TCase stands for Case Temperature and is the maximum temperature allowed at the processor Integrated Heat Spreader (IHS).

The myth that analysis programs read the wrong temperature, such as HWmonitor, AIDA64, Core Temp, etc., does not hold, either, as I am a user of an FX-6300, and if I compare AMD Overdrive (taking into account Tcase 70 ° C) with the analysis programs, the temperature is exactly the same.
If your motherboard has a diode touching the cpu then keep it under 70c. You will throttle if the temps get too high as you hit Tj Max. Tjunction max is the temperature at which transistor functionality begins to break down (become unreliable).

Tcase is pretty much what the temp of your CPU is, measured by your motherboard, I remember a little diode touching the bottom of the cpu in one of my old systems (another way is to touch the pins in the cpu socket. The Tjunction max, or Thermal Junction max is the maximum temperature your CPU is rated for. That's what the proprietary-unit thermal-margin scale used by AMD's overdrive software.
 
Last edited:
Jan 28, 2020
91
8
45
2
Are you thermal throttling? Stock coolers are mostly about getting the job done and not about keeping temps to a minimum. Also if you have been running the system for years, have you cleaned out the dust from the heatsink and fan?

The max Boost Clock for the FX 6300 is up to 3.8GHz AMD. http://www.cpu-world.com/CPUs/Bulldozer/AMD-FX-Series FX-6300.html shows,

Boosted P states [1]#1: 4100 MHz, 1.425V
#2: 3800 MHz, 1.4125V

So your 4.1GHz @ 1.425 volts is normal.
Maximum operating temperature ?70.5°C

You can check your thermal margin as well if you like.
The Tcase Temperature of FX 6300 processor is 70ºC. source email from amd TCase stands for Case Temperature and is the maximum temperature allowed at the processor Integrated Heat Spreader (IHS).



If your motherboard has a diode touching the cpu then keep it under 70c. You will throttle if the temps get too high.
Yes it seems to thermal throttle as far as I know. Also I fully disassembled the PC few days back and cleaned every nook and corner.
As I observed under a 15-20 mins of stress test, the CPU hits the boost to 4.118GHz on 1.42V, and immediately dips to stock speeds 3.5-3.8GHz somewhere near 1.375 to 1.4V.
The temps recorded were around 64C from the motherboard sensor and around 62C from the package.
This is the reason I thought of undervolting the chip, but it seems the BIOS doesn't let me undervolt under 1.3V. Whereas I do have an option to undervolt in AOD which I wouldn't like to try.
 

zx128k

Upstanding
Nov 23, 2019
474
88
290
6
Yes it seems to thermal throttle as far as I know. Also I fully disassembled the PC few days back and cleaned every nook and corner.
As I observed under a 15-20 mins of stress test, the CPU hits the boost to 4.118GHz on 1.42V, and immediately dips to stock speeds 3.5-3.8GHz somewhere near 1.375 to 1.4V.
The temps recorded were around 64C from the motherboard sensor and around 62C from the package.
This is the reason I thought of undervolting the chip, but it seems the BIOS doesn't let me undervolt under 1.3V. Whereas I do have an option to undervolt in AOD which I wouldn't like to try.
That sounds correct, you have a heavy load, cpu boosts to the top P state but temps rise fast. So starts dropping P states so temps are stable. If you hit TJ max your cpu will really pull on the brakes. Stock frequency of 3500 MHz is likely normal for a heavy load and stock cooler. 3800 and 4.1 Ghz are boost frequencies.

The low power states are:
Low power P states [1]#1: 3000 MHz, 1.225V
#2: 2500 MHz, 1.125V
#3: 2000 MHz, 1.025V
#4: 1400 MHz, 0.9V

Here in this video a stock 6300 with a stock cooler, cycles between 3500MHz and 3000MHz in Prime 95.


Also in Aida64 3500MHz


What you are stating above appears to be stock operation.
 
Jan 28, 2020
91
8
45
2
That sounds correct, you have a heavy load, cpu boosts to the top P state but temps rise fast. So starts dropping P states so temps are stable. If you hit TJ max your cpu will really pull on the brakes. Stock frequency of 3500 MHz is likely normal for a heavy load and stock cooler. 3800 and 4.1 Ghz are boost frequencies.

The low power states are:
Low power P states [1]#1: 3000 MHz, 1.225V
#2: 2500 MHz, 1.125V
#3: 2000 MHz, 1.025V
#4: 1400 MHz, 0.9V

Here in this video a stock 6300 with a stock cooler, cycles between 3500MHz and 3000MHz in Prime 95.


Also in Aida64 3500MHz


What you are stating above appears to be stock operation.
But my temps on stress test are more than what is in both the videos. Maybe I should do a stress test for around 10 mins and post an update here about it with the screenshots. Do you recommend that I bump the voltage from 1.3 to 1.31 or 1.32V and do a tiny OC to have stable clock speeds?
 

zx128k

Upstanding
Nov 23, 2019
474
88
290
6
From what I can see the core clock is 3800MHz and the max is the 4100MHz. So you are boosting normally.

You have 2 separate CPU temp readings on most AM3+ systems.
  1. A "package temp" reported by the CPU itself, which is actually not a "temperature" reading at all, rather, is a proprietary-unit thermal-margin scale.
  2. A "socket temp" reported by the SuperIO chip on your motherboard, from a probe under the socket, or near the socket on the motherboard. https://pcpartpicker.com/forums/topic/79614-hyper-evo-212-amd-fx-6300-temperatures-i-am-a-newb
The diode temp is the temperature of the cpu itself and should be close to the Tcase. The thermal-margin is the distance to the throttling/shut down temperature. The link I give above speculates how to turn this into a temperature and implies the result is accurate. link to source, use google translate

One temperature reading is the diode on the motherboard and this is TMPIN0 in HWmonitor source 1 Looks like you're right because I just put in a thermal probe from a fan controller I have and it's within 2C of temps reported by TMPIN0.

TMPIN0 is 63c which means you are close to 70c. This should be close to the current tcase value. Max tcase being 70c. link to source, use google translate

You can also check the AMD overdrive software and see your thermal-margin just to be complete about it. Note this software has been discontinued. https://www.amd.com/en/technologies/amd-overdrive A link still exists on other websites. https://www.techspot.com/downloads/4645-amd-overdrive.html Last updated: March 1, 2018

If the overdrive software does not work on your system you will have to use only the motherboard diode. You will be able to tell you are too hot when the cpu really throttles a lot (below 3500MHz) and if you reach the shut down temps your PC will turn off. The diode should be close to the tcase value so you should see it increase towards 70c and at some point the cpu should take action. Lower temps will likely give you a higher more stable boost.


source
 
Last edited:
Jan 28, 2020
91
8
45
2
From what I can see the core clock is 3800MHz and the max is the 4100MHz. So you are boosting normally.

You have 2 separate CPU temp readings on most AM3+ systems.
  1. A "package temp" reported by the CPU itself, which is actually not a "temperature" reading at all, rather, is a proprietary-unit thermal-margin scale.
  2. A "socket temp" reported by the SuperIO chip on your motherboard, from a probe under the socket, or near the socket on the motherboard. https://pcpartpicker.com/forums/topic/79614-hyper-evo-212-amd-fx-6300-temperatures-i-am-a-newb
The diode temp is the temperature of the cpu itself and should be close to the Tcase. The thermal-margin is the distance to the throttling/shut down temperature. The link I give above speculates how to turn this into a temperature and implies the result is accurate. link to source, use google translate

One temperature reading is the diode on the motherboard and this is TMPIN0 in HWmonitor source 1 Looks like you're right because I just put in a thermal probe from a fan controller I have and it's within 2C of temps reported by TMPIN0.

TMPIN0 is 63c which means you are close to 70c. This should be close to the current tcase value. Max tcase being 70c. link to source, use google translate

You can also check the AMD overdrive software and see your thermal-margin just to be complete about it. Note this software has been discontinued. https://www.amd.com/en/technologies/amd-overdrive A link still exists on other websites. https://www.techspot.com/downloads/4645-amd-overdrive.html Last updated: March 1, 2018

If the overdrive software does not work on your system you will have to use only the motherboard diode. You will be able to tell you are too hot when the cpu really throttles a lot (below 3500MHz) and if you reach the shut down temps your PC will turn off. The diode should be close to the tcase value so you should see it increase towards 70c and at some point the cpu should take action. Lower temps will likely give you a higher more stable boost.


source
Well I did a stress test with AMD Turbo Boost turned off. I got these thermal margin in AOD and results from HWInfo. Seems like the tjmax is about 72C with the thermal margin added to the temps from the package. So what are your thoughts on this? Do I have some headroom to OC and still maintain optimal temps?
 

zx128k

Upstanding
Nov 23, 2019
474
88
290
6
Well I did a stress test with AMD Turbo Boost turned off. I got these thermal margin in AOD and results from HWInfo. Seems like the tjmax is about 72C with the thermal margin added to the temps from the package. So what are your thoughts on this? Do I have some headroom to OC and still maintain optimal temps?
You could get 4GHz but you will have to watch the temps. I am not sure if you can without overheating.
 
Jan 28, 2020
91
8
45
2
You could get 4GHz but you will have to watch the temps. I am not sure if you can without overheating.
Will try overclocking to 4GHz or even 3.8GHz and watch the temps accordingly and stick to one of the oc from the two and post the updates. Thank you for the replies and for your precious time, I really appreciate your help.
 
Jan 28, 2020
91
8
45
2
You could get 4GHz but you will have to watch the temps. I am not sure if you can without overheating.
These are the results and temperatues by changing these settings in the BIOS:
  • Increased CPU/HT Reference to about 5%
  • Decreased HT Clock to 2100MHz to match the NB Clock and did a stress test for about 10-12 mins each.
Btw everything's running on default voltages for CPU, NB, HT and MEM.

Stock/Default multiplier x17.5(3500MHz) 3700MHz
Multiplier changed to x18.0(3600MHz) 3800MHz
 
Last edited:

zx128k

Upstanding
Nov 23, 2019
474
88
290
6
These are the results and temperatues by changing these settings in the BIOS:
  • Increased CPU/HT Reference to about 5%
  • Decreased HT Clock to 2100MHz to match the NB Clock and did a stress test for about 10-12 mins each.
Btw everything's running on default voltages for CPU, NB, HT and MEM.

Stock/Default multiplier x17.5(3500MHz) 3700MHz
Multiplier changed to x18.0(3600MHz) 3800MHz
Looks good, also AMD overdrive thermal margin can be helpful to know as well.
 
Jan 28, 2020
91
8
45
2
Looks good, also AMD overdrive thermal margin can be helpful to know as well.
Oh sorry I forgot to take screenshots of AOD. I remember the thermal margins was hovering around 16.5-18C on stock multiplier 3700MHz and 14.5-16C on x18(3600MHz) multiplier 3800MHz. Are these temperatures good?? Do I need to look after the VRM's temperatures even though I'm running on stock 1.2v Vcore???
 

zx128k

Upstanding
Nov 23, 2019
474
88
290
6
Oh sorry I forgot to take screenshots of AOD. I remember the thermal margins was hovering around 16.5-18C on stock multiplier 3700MHz and 14.5-16C on x18(3600MHz) multiplier 3800MHz. Are these temperatures good?? Do I need to look after the VRM's temperatures even though I'm running on stock 1.2v Vcore???
You always check VRM temperatures, it's a paranoid habit under high loads. The power dyn needed goes up with frequency, capacitance and voltage squared. If you increase the frequency then the power cpu will increase which will mean more load on the VRM's.

Thermal Margin indicates how far the current operating temperature is below the maximum operating temperature of the processor. So 14.5-16c would be okay.
 
Jan 28, 2020
91
8
45
2
You always check VRM temperatures, it's a paranoid habit under high loads. The power dyn needed goes up with frequency, capacitance and voltage squared. If you increase the frequency then the power cpu will increase which will mean more load on the VRM's.

Thermal Margin indicates how far the current operating temperature is below the maximum operating temperature of the processor. So 14.5-16c would be okay.
Yeah, the thermal margins are above 10C which is said to be okay, but I'm just concerned about the overclock affecting VRM's, NB and the RAM die's which none of them have a heatsink on them. And hoping that they just don't give up on me anytime until my future upgrade.
 

zx128k

Upstanding
Nov 23, 2019
474
88
290
6
Yeah, the thermal margins are above 10C which is said to be okay, but I'm just concerned about the overclock affecting VRM's, NB and the RAM die's which none of them have a heatsink on them. And hoping that they just don't give up on me anytime until my future upgrade.
HWiNFO normally can read the VRM temps and the NB. RAM you have to read with a Infrared thermometer if you have one. Fingers are not really a good idea. RAM heatsinks don't really do anything. RAM keeping the voltage safe is important. DDR3 1600 should be fine at stock, unless you had to overclock them and increase the voltage. That said good air flow should be enough.

Overclocking is a risk on an old system but we all do it.

You need a good buffer for cpu temps because in winter ambient temps are more forgiving.
 
Jan 28, 2020
91
8
45
2
HWiNFO normally can read the VRM temps and the NB. RAM you have to read with a Infrared thermometer if you have one. Fingers are not really a good idea. RAM heatsinks don't really do anything. RAM keeping the voltage safe is important. DDR3 1600 should be fine at stock, unless you had to overclock them and increase the voltage. That said good air flow should be enough.

Overclocking is a risk on an old system but we all do it.

You need a good buffer for cpu temps because in winter ambient temps are more forgiving.
But there's no sensors for VRM's and NB on this board. That's the problem I have to face with :( And coming to RAM it is just a 80MHz bump on stock voltage, which I think is safe!!! Idk yet. Also coming to airflow, the system's got 2 intakes and an exhaust.
I really want some bump in gaming performance, yet I'm confused(or even scared) about the overclock. But man!!! You are such a good person for lending your time and helping me out through all this fuss. I really am very grateful to you :)
 
Does your mobo have a core performance boost multiplier? ( Turbo Boost )

If so you could increase the voltage slightly and increase the cpb multiplier for it to boost higher.

My wifes 6300 on a gigabyte 970 SLi
Has it and I have it set to around 23.5 to 24.5 and it turbos upto 4.822ghz.
I can start unigene heaven benchmark and by the time it is done all cores have hit the 4.822ghz high by watching hwinfo64.

Then you wouldn't need to be so worried about the temps and voltages.
If you would like I can and will send pics when I get back home thur evening.
 
Jan 28, 2020
91
8
45
2
Does your mobo have a core performance boost multiplier? ( Turbo Boost )

If so you could increase the voltage slightly and increase the cpb multiplier for it to boost higher.

My wifes 6300 on a gigabyte 970 SLi
Has it and I have it set to around 23.5 to 24.5 and it turbos upto 4.822ghz.
I can start unigene heaven benchmark and by the time it is done all cores have hit the 4.822ghz high by watching hwinfo64.

Then you wouldn't need to be so worried about the temps and voltages.
If you would like I can and will sent pics when I get back home thur evening.
Yes the board has both Turbo Boost and CPU multipliers. Right now I've disabled Turbo Boost, and just increased CPU/HT Reference clock to about 10% (from 200-210) which changed RAM freq to 840MHz from 800MHz, NB to 2100MHz from 2000MHz and changed the multiplier to x18.0 which is equivalent to 3.8GHz running on stock voltages. Here are the results and temps over 10-15mins of p95 stress test.

Stock/Default multiplier x17.5(3500MHz) 3700MHz
Multiplier changed to x18.0(3600MHz) 3800MHz

The mobo has no heatsinks on VRM's and RAM's, no sensors to check the temps for NB, VRM's and that's a concern for me.
 

ASK THE COMMUNITY

TRENDING THREADS