Needs higher voltage for increased performance
Needs higher voltage for increased performance
I recently purchased my PC and put it together myself. Most of my tasks require intense processing power (3D rendering and graphics), so I increased the CPU speed to improve performance.
PC Specifications
Motherboard: Asus X99 Deluxe II
Cooler: Thermaltake RGB 360 Triple Rad Watercooler
CPU: i7 6900k (overclocked to 3.2 GHz at 1.21V)
GPU: Asus ROG STRIX 1080 (x2)
Ram: Corsair Vengeance DDR4-3000 128MB
SSD: 1TB M.2 960 EVO
Overclocking Adjustments (changes made from standard settings):
Overclock Tuner: XMP
CPU Strap: 100MHz
CPU: Sync All Cores 42
Min/Max CPU Cache Ratio: 30
DRAM Frequency: 3000MHz
CPU Voltage: Adaptive with negative offset (other parameters adjusted automatically; stabilized at 1.15V per HW monitor)
All adjustments were made slowly and step by step before reaching 4.2 GHz. During the first month, everything remained stable. I bought the components on the second of the year and assembled them that week, following 12-hour Aida64 and Prime95 benchmarks. Beginning last week, I started experiencing BSODs (often with code WHEA_UNCORRECTABLE_ERROR) when the CPU was under heavy load. Each time a BSOD occurred, I reduced the negative offset on the CPU voltage until it reached 0. The frequency of these crashes increased, and now the system won’t boot at the overclocked specs even with 0 offset on adaptive voltage. Currently, the only way to start is by manually resetting CMOS and reverting to default settings.
What went wrong? Did I damage a valuable component or did I install a faulty processor? Are there any warning signs in my configuration or overclocking details?
Thank you very much for your assistance.
Sure, let's try again with a fresh approach. Let's go through some troubleshooting steps. Please be patient as we'll need time to locate the issue. Bring your BIOS back to its default settings (clear CMOS or use the original BIOS) and check if Windows encounters any problems.
Instead of a long explanation, here are the possible solutions for Kernel_Mode_Trap_Error:
[Link to troubleshooting guide]
WHEA errors usually relate to the CPU, but may also involve BIOS or other hardware components like USB devices or PCI cards. This can be time-consuming to diagnose. If the issue remains unresolved, a comprehensive analysis of your system's BSOD is necessary. Let's proceed and see what we find.
There are several...
Hello WlilJS
Please verify that your Bios is current with version 1401 for improved system performance. Offsets are not necessary until version 4.6GH, and even then they might not be needed based on stability. Switch your AI Overclock Tuner to Manual mode instead of XMP for more precise control. Adjust the Core Voltage to 1.375V and set the frequency to 4.6GHz. Navigate to Digi+ Power Control, configure the LLC setting to 8 with CPU Current capacity at 120%. Turn off all sleep states such as C1E, as they can interfere with overclocking.
On your ASUS EPU, ensure that "Disabled" is selected. Set the CPU to Sync All Cores at 46. Proceed to Internal CPU Power Management and disable Speed Step and Turbo mode. Keep other settings on Auto for now. Save by pressing F10 and exit.
After installation, try to boot into Windows. If your overclock is accepted, perform a stress test. Avoid using P95 as it can be too demanding. Using AIDA64 along with CPUID HWMonitor on your desktop will give you all the necessary information and testing tools.
In AIDA64, select only CPU, FPU, and Cache options. Start by testing for 10 minutes to verify PSU rail voltages and detect any voltage drops. This test also helps assess the effectiveness of your cooling solution.
If you need analysis, capture screenshots at the 10-minute mark and share them using your preferred file host. Upload the images here, and I’ll provide the BBurl link for further evaluation.
MeanMachine41 :
Check your BIOS for updates to version 1401 to ensure optimal system performance. Offsets aren't necessary until version 4.6GH, and even then they might not be needed based on stability. Switch your AI Overclock Tuner to Manual mode instead of XMP for more precise control. Adjust the Core Voltage to 1.375V and set the frequency to 4.6GHz. Navigate to Digi+ Power Control, configure LLC to 8 with CPU Current at 120%, and turn off all Sleep states such as C1E. ASUS EPU is now disabled.
CPU: Sync All Cores to 46. Proceed to Internal CPU Power Management and disable Speed Step and Turbo mode. Keep the remaining settings on Auto for now. Save with F10 and exit.
Hopefully, the system boots correctly to Windows after your OC is accepted. After that, perform a stress test. Avoid using P95 as it's too demanding. Using AIDA64 alongside CPUID HWMonitor on your desktop will give you all the necessary information and testing tools.
In AIDA64, select only CPU, FPU, and Cache options. Start with a 10-minute test to verify PSU rail voltages and detect any voltage drops. This will also confirm if your cooling system is functioning properly.
If you need analysis of your results, take screenshots at the 10-minute mark and share them using your preferred file host. Upload the images here, and I’ll provide the BBurl link for further evaluation.
A few follow-up questions: Are there any risks in increasing the voltage and core clock to such high levels? I thought 4.2 GHz was quite aggressive, and I haven’t seen a 6900k CPU exceed 4.4 GHz. Can you confirm you’re not mixing up a 6700k with a 6900k?
Also, please be certain this isn’t a mistake before pushing your CPU to that limit, as I can’t afford another failure.
If it can't accept the OC then the system would reject it. If you prefer a gradual approach, begin testing from 4.3-4.4GHz with 1.300V (even turbo mode works). The outcome relies on the chip quality; some users have achieved up to 4.6GHz. At 1.375V on the Core, no damage occurs. During stress tests in AIDA64, the CPU may overheat—stop if temperatures reach 70°C. If temperatures remain stable below 70°C, the chip is solid and the cooling system is effective. OC performance under load shows consistent high temperatures, which can affect the CPU. My recommendations are based on the AIO Thermaltake RGB 360 Triple Rad Watercooler.
MeanMachine41 :
If the system can't handle the OC then it will likely reject it. If you prefer testing in phases, start with 4.3-4.4GHz and 1.300V (even turbo mode works). It all comes down to the chip quality; some users have reached 4.6GHz. At 1.375V on the Core, no damage should occur. The CPU may overheat during stress testing in AIDA64—stop if temperatures hit 70°C. If temperatures remain stable below 70°C, you’ve got a solid chip and your cooling is effective.
OCing consistently high temps under load can hurt the CPU. My recommendations are based on the AIO Thermaltake RGB 360 Triple Rad Watercooler. I followed exactly what you suggested to overclock. The PC accepted 1.375V at 4.4GHz. I was able to boot and start HWmonitor before issues arose. It reported a default clock speed of 3.2GHz before the system crashed completely. I had to reset the MBoard again to recover.
If we dig into why this is happening, it might be better than jumping into overclocking. As mentioned in my first post, I encountered errors like Kernel_Mode_Trap_Error and Whea_Uncorrectable_Error at much lower overclocks than what you’re proposing.
Running the "check system" on my PSU (Corsair AZ1200i) gave a green light. All components are under a month old. Apart from taking everything out and resting it, I’m not sure how to fix this. My BIOS is also up to date.
Ok WlilJS, let's begin anew and tackle some troubleshooting steps. Please be patient as this might take a while to identify the issue.
Return your BIOS settings to the default (clear CMOS or use the original BIOS) and observe if Windows encounters any problems.
Instead of a long explanation, here are solutions for Kernel_Mode_Trap_Error:
Visit the provided link for guidance on fixing unexpected kernel traps on Windows 10.
WHEA errors usually point to CPU issues, but they can also stem from BIOS or hardware (including USB devices or PCI cards), which may require more time to diagnose. If the problem persists, a thorough BSOD analysis will be necessary. Let's proceed and see what we discover first.
We can perform several checks to confirm the issue isn't related to the operating system.
In an elevated command prompt, run "sfc /scannow" without quotes to verify your OS files' integrity.
Inspect Device Manager for any yellow triangles suggesting a driver problem.
Utilize Event Viewer to look for persistent critical events.
Execute "chkdsk" to assess the health of your operating system disk.
Run a virus or malware scanner to rule out other causes.
If everything appears normal, ensure your DIMM modules are functioning correctly. I use Memtest86 from a USB drive—perform the test on each module and then collectively for three passes to confirm.
Please let me know your findings before we consider further steps.