F5F Stay Refreshed Power Users Overclocking I can assist you with overclocking your QX6850.

I can assist you with overclocking your QX6850.

I can assist you with overclocking your QX6850.

Pages (2): 1 2 Next
B
BernyTheMan
Member
180
09-15-2016, 07:30 AM
#1
ok so i have a QX6850 that is liquid cooled via an AIO. temps idle between 37-40c. at max load in prime95 in blend test at stock speeds 3.0GHz temps max out around 56-57c. while not new to overclocking my previous chip was a core 2 duo E8500 im new to overclocking quads so i have a few questions
#1 what temps should i watch out for while stress testing in prime i.e. keep CPU under X temp/temps when stressing for stable speeds
my goal is to reach 4.0GHz stable 24/7
#2 aside from reading the overclocking guide on here is there anything else i should know ABOUT MY QX6850 IN TERMS OF OVERCLOCKING I KNOW THESE THINGS ARE KNOWN TO BE HEATERS
lastly should i test under small FFT's or blend wanna make sure both cpu and ram are stable
also what options do i disable in the bios so while stress testing the cpu will run full speed all the time and not lower its multi and stuff
BEFORE I FORGET im using MX-4 as my thermal paste below is a link to my AIO just incase
http://www.coolermaster.com/cooling/...quid-lite-240/
AIO is run in push pull setup because i had the extra fans lying around
B
BernyTheMan
09-15-2016, 07:30 AM #1

ok so i have a QX6850 that is liquid cooled via an AIO. temps idle between 37-40c. at max load in prime95 in blend test at stock speeds 3.0GHz temps max out around 56-57c. while not new to overclocking my previous chip was a core 2 duo E8500 im new to overclocking quads so i have a few questions
#1 what temps should i watch out for while stress testing in prime i.e. keep CPU under X temp/temps when stressing for stable speeds
my goal is to reach 4.0GHz stable 24/7
#2 aside from reading the overclocking guide on here is there anything else i should know ABOUT MY QX6850 IN TERMS OF OVERCLOCKING I KNOW THESE THINGS ARE KNOWN TO BE HEATERS
lastly should i test under small FFT's or blend wanna make sure both cpu and ram are stable
also what options do i disable in the bios so while stress testing the cpu will run full speed all the time and not lower its multi and stuff
BEFORE I FORGET im using MX-4 as my thermal paste below is a link to my AIO just incase
http://www.coolermaster.com/cooling/...quid-lite-240/
AIO is run in push pull setup because i had the extra fans lying around

S
soul_harveste
Member
223
09-15-2016, 02:19 PM
#2
Keep your NB voltage and temperatures in check; increasing the voltage improves stability but also generates more heat as stress increases. On the other hand, raising the CPU voltage can help with lower NB voltages, though it would also introduce additional heat from the CPU. The results will depend on the specific board you're using. I tested it with the Asus Rampage Extreme(x48) to gain practical insights. Make sure you stay below the processor's maximum temperature threshold. Your link isn't functional for me, so I can't confirm your cooling setup. You can review my AMS build log (under my signature) to see how I approached the overclocking process.
S
soul_harveste
09-15-2016, 02:19 PM #2

Keep your NB voltage and temperatures in check; increasing the voltage improves stability but also generates more heat as stress increases. On the other hand, raising the CPU voltage can help with lower NB voltages, though it would also introduce additional heat from the CPU. The results will depend on the specific board you're using. I tested it with the Asus Rampage Extreme(x48) to gain practical insights. Make sure you stay below the processor's maximum temperature threshold. Your link isn't functional for me, so I can't confirm your cooling setup. You can review my AMS build log (under my signature) to see how I approached the overclocking process.

C
55
09-15-2016, 08:39 PM
#3
Keep your NB voltage and temperatures in check; increasing the voltage improves stability but also generates more heat as stress increases. On the other hand, raising the CPU voltage can help with lower NB voltages, though it would also introduce additional heat from the CPU. The results will depend on the specific board you're using. I tested it with the Asus Rampage Extreme(x48) to gain practical insights. Make sure you stay below the processor's maximum temperature threshold. Your link isn't functional for me, so I can't confirm your cooling setup. You can review my AMS build log (under my signature) to see how I approached the overclocking process.
C
carebearisboss
09-15-2016, 08:39 PM #3

Keep your NB voltage and temperatures in check; increasing the voltage improves stability but also generates more heat as stress increases. On the other hand, raising the CPU voltage can help with lower NB voltages, though it would also introduce additional heat from the CPU. The results will depend on the specific board you're using. I tested it with the Asus Rampage Extreme(x48) to gain practical insights. Make sure you stay below the processor's maximum temperature threshold. Your link isn't functional for me, so I can't confirm your cooling setup. You can review my AMS build log (under my signature) to see how I approached the overclocking process.

L
LumpVersusMC
Junior Member
26
09-19-2016, 04:22 PM
#4
Keep your NB voltage and temperatures in check; increasing the voltage improves stability but also generates more heat. On the other hand, raising the CPU voltage can help with lower NB voltages, though it would also require managing additional CPU heat. The results will depend on the specific board you're using. I tested it with the Asus Rampage Extreme(x48) to gain practical insights. Make sure your processor isn't hitting its maximum temperature limit before proceeding. Your link isn't functional for me, so I can't confirm the cooling setup. You can review my AMS build log (under my signature) for details on the overclock process. This is the board I'm currently working with: https://www.gigabyte.com/Motherboard/GA-...-rev-10#sp
L
LumpVersusMC
09-19-2016, 04:22 PM #4

Keep your NB voltage and temperatures in check; increasing the voltage improves stability but also generates more heat. On the other hand, raising the CPU voltage can help with lower NB voltages, though it would also require managing additional CPU heat. The results will depend on the specific board you're using. I tested it with the Asus Rampage Extreme(x48) to gain practical insights. Make sure your processor isn't hitting its maximum temperature limit before proceeding. Your link isn't functional for me, so I can't confirm the cooling setup. You can review my AMS build log (under my signature) for details on the overclock process. This is the board I'm currently working with: https://www.gigabyte.com/Motherboard/GA-...-rev-10#sp

T
TntXDgames
Junior Member
12
09-20-2016, 10:48 PM
#5
I managed to get my QX6800 to run at 4GHz on a locked BIOS configuration with the CPUZ in my notes and a Dell heatpipe cooler. It didn’t work well there. With an unlocked CPU, the process was straightforward. The 65nm processors can tolerate some voltage adjustments. Given your board already features strong VRM and chipset cooling, I recommend increasing the voltage to 1.5875V and setting the multiplier to 12x333. You can adjust this in Windows; if it crashes, it will automatically reboot to normal speed and voltage. Then gradually increase or decrease the voltage. Using Throttlestop 6.00 software lets you manage voltage and multiplier during testing. Find your settings and consider locking the overclock in the BIOS if needed. You might also try raising the FSB frequency to 1600 in BIOS and start at 3.6GHz with a 10x400 test. There’s a discussion on this approach here: https://www.techpowerup.com/forums/threa...cs.235975/
T
TntXDgames
09-20-2016, 10:48 PM #5

I managed to get my QX6800 to run at 4GHz on a locked BIOS configuration with the CPUZ in my notes and a Dell heatpipe cooler. It didn’t work well there. With an unlocked CPU, the process was straightforward. The 65nm processors can tolerate some voltage adjustments. Given your board already features strong VRM and chipset cooling, I recommend increasing the voltage to 1.5875V and setting the multiplier to 12x333. You can adjust this in Windows; if it crashes, it will automatically reboot to normal speed and voltage. Then gradually increase or decrease the voltage. Using Throttlestop 6.00 software lets you manage voltage and multiplier during testing. Find your settings and consider locking the overclock in the BIOS if needed. You might also try raising the FSB frequency to 1600 in BIOS and start at 3.6GHz with a 10x400 test. There’s a discussion on this approach here: https://www.techpowerup.com/forums/threa...cs.235975/

L
Lewiswilson4
Member
148
09-20-2016, 11:31 PM
#6
There are numerous multiplier overclocks available on the chart, including the two I recommended.
L
Lewiswilson4
09-20-2016, 11:31 PM #6

There are numerous multiplier overclocks available on the chart, including the two I recommended.

D
dehunter456
Member
237
09-24-2016, 01:11 AM
#7
william p :
I tried running my QX6800 at 4GHz on a locked BIOS Dell system (CPUZ in my notes) with a heatpipe cooler. It didn’t work well there. But with an unlocked CPU, it was straightforward. Those 65nm chips can manage some voltage adjustments. Given your board already has solid VRM and chipset cooling, I recommend increasing the voltage to 1.5875V and setting the multiplier to 12x333. You should be able to do this in Windows; if it crashes, it will simply restart at normal speed/voltage. Then gradually lower or raise the voltage again. If you use Throttlestop 6.00, you can control voltage and multiplier during operation. Check your settings and consider locking the overclock in BIOS if needed. You might also try raising FSB to 1600 in BIOS and start at 3.6GHz with a 10x400 test.

Here’s a relevant discussion:
https://www.techpowerup.com/forums/threa...cs.235975/
If your cooling is adequate, reaching 4GHz should take about 10 minutes.

4ghz wasn’t an option even with my 240mm AIO water cooling, as it would hit the 80s at Clost and only drop to 3.6GHz with MX-4 thermal paste.
D
dehunter456
09-24-2016, 01:11 AM #7

william p :
I tried running my QX6800 at 4GHz on a locked BIOS Dell system (CPUZ in my notes) with a heatpipe cooler. It didn’t work well there. But with an unlocked CPU, it was straightforward. Those 65nm chips can manage some voltage adjustments. Given your board already has solid VRM and chipset cooling, I recommend increasing the voltage to 1.5875V and setting the multiplier to 12x333. You should be able to do this in Windows; if it crashes, it will simply restart at normal speed/voltage. Then gradually lower or raise the voltage again. If you use Throttlestop 6.00, you can control voltage and multiplier during operation. Check your settings and consider locking the overclock in BIOS if needed. You might also try raising FSB to 1600 in BIOS and start at 3.6GHz with a 10x400 test.

Here’s a relevant discussion:
https://www.techpowerup.com/forums/threa...cs.235975/
If your cooling is adequate, reaching 4GHz should take about 10 minutes.

4ghz wasn’t an option even with my 240mm AIO water cooling, as it would hit the 80s at Clost and only drop to 3.6GHz with MX-4 thermal paste.

X
Xtun3r
Member
154
10-03-2016, 02:04 PM
#8
there are numerous multiplier overclocks listed on the chart, including the two I recommended.
here is a link to my cpu cooler: http://www.coolermaster.com/cooling/cpu-...-lite-240/
the other one can be found at http://valid.x86.fr/top-cpu/496e74656c28...303047487a
X
Xtun3r
10-03-2016, 02:04 PM #8

there are numerous multiplier overclocks listed on the chart, including the two I recommended.
here is a link to my cpu cooler: http://www.coolermaster.com/cooling/cpu-...-lite-240/
the other one can be found at http://valid.x86.fr/top-cpu/496e74656c28...303047487a

J
jack1821
Member
64
10-03-2016, 06:06 PM
#9
As I said no stability there (4GHz). I unplugged the PWM wire form my 2 motor fan, got a validation and shut it down. I was gaming at 3.72Ghz. Unfortunately the QX6850 has full multiplier steps like the other 65nm extremes. You might need to use SetFSB to turn down the FSB so you can add another multiplier, and then tune back up with SetFSB to 3.7-3.8GHz.
For example 13X300=3900 MHz/3.9GHz
I list 3.72GHz in my sig. but the validation link is for 4Ghz.
Temps in prime will be higher than normal use. I run 3x Prime 95 which simulates useful gaming loads since 100% CPU = bottleneck. 75% is useful load to represent gaming.
To get 4GHz on that CPU Dell used an H2C Peltier asisited water loop in their XPS 720/730 computers.
J
jack1821
10-03-2016, 06:06 PM #9

As I said no stability there (4GHz). I unplugged the PWM wire form my 2 motor fan, got a validation and shut it down. I was gaming at 3.72Ghz. Unfortunately the QX6850 has full multiplier steps like the other 65nm extremes. You might need to use SetFSB to turn down the FSB so you can add another multiplier, and then tune back up with SetFSB to 3.7-3.8GHz.
For example 13X300=3900 MHz/3.9GHz
I list 3.72GHz in my sig. but the validation link is for 4Ghz.
Temps in prime will be higher than normal use. I run 3x Prime 95 which simulates useful gaming loads since 100% CPU = bottleneck. 75% is useful load to represent gaming.
To get 4GHz on that CPU Dell used an H2C Peltier asisited water loop in their XPS 720/730 computers.

H
Hazardrider01
Junior Member
39
10-03-2016, 06:41 PM
#10
I had a task I needed to complete. You can find the link in my profile for ideas on overclocking Dell BTX computers.
The MB isn't specified.
CPUZ for QX6850
http://valid.x86.fr/top-cpu/496e74656c28...303047487a
CPUZ for QX6800 displaying multiplier overclocks 15x266 fsb. My model is #5, it was #2 when I shared it.
You should begin by addressing the CPU heat spreader. Use progressively finer sandpaper on glass to achieve a better metal-to-metal contact with the heatsink, which should already be polished if quality exists. Investigate which TIM compound is optimal at the moment. Apply only what's necessary. Aim for the thinnest layer possible, avoiding trapped air—ideally a single small dot in the center that expels air as it spreads.
Consider VRM cooling if it isn't present. Ensure airflow enters through the cooler and exits via the case, covering the VRMs too. Be mindful that air doesn’t exit straight from fans; it spreads in a cone shape with a significant gap in the middle unless restricted. Many small adjustments contribute gradually toward reaching 4GHz with these processors. Water cooling is ineffective below room temperature. Increasing RT air flow through the radiator will help, but heat must first be expelled from the CPU, and the power supply must deliver sufficient watts for these older units.
You might consider an LGA775 Club socket at OCN—they really like those.
http://www.overclock.net/forum/8-intel-g...-club.html
H
Hazardrider01
10-03-2016, 06:41 PM #10

I had a task I needed to complete. You can find the link in my profile for ideas on overclocking Dell BTX computers.
The MB isn't specified.
CPUZ for QX6850
http://valid.x86.fr/top-cpu/496e74656c28...303047487a
CPUZ for QX6800 displaying multiplier overclocks 15x266 fsb. My model is #5, it was #2 when I shared it.
You should begin by addressing the CPU heat spreader. Use progressively finer sandpaper on glass to achieve a better metal-to-metal contact with the heatsink, which should already be polished if quality exists. Investigate which TIM compound is optimal at the moment. Apply only what's necessary. Aim for the thinnest layer possible, avoiding trapped air—ideally a single small dot in the center that expels air as it spreads.
Consider VRM cooling if it isn't present. Ensure airflow enters through the cooler and exits via the case, covering the VRMs too. Be mindful that air doesn’t exit straight from fans; it spreads in a cone shape with a significant gap in the middle unless restricted. Many small adjustments contribute gradually toward reaching 4GHz with these processors. Water cooling is ineffective below room temperature. Increasing RT air flow through the radiator will help, but heat must first be expelled from the CPU, and the power supply must deliver sufficient watts for these older units.
You might consider an LGA775 Club socket at OCN—they really like those.
http://www.overclock.net/forum/8-intel-g...-club.html

Pages (2): 1 2 Next