9700K Overclock
9700K Overclock
Guys,
Please go through this thoroughly with me. I used to perform overclocking on a K6-2 500 a long time ago, but I’m now getting updates with new settings.
I own a Maximus XI Hero (Asus Rog) and a 9700k. I recently achieved an overclock to 4.8 GHz by simply adjusting the default BIOS options:
- Sync all cores to 48x multiplier
- Load Line Calibration (LLC) set to 3
I ran the official Intel app and completed a burn-in test lasting 30 minutes (https://ibb.co/mqCR3JT). The results were promising (see the attached image):
The clock speed remained stable at 4.8 GHz
The hottest core reached 89 degrees
Voltage: 1.3345v
Memory frequency: 99.98 MHz
I have a few questions:
- Is that temperature normal? I know less than 85°C is ideal, but the test was quite aggressive. How much is acceptable for a burn-in test?
- Intel Boost stopped working (it doesn’t matter to me now—just want to confirm it’s normal)
- The VID says less is better, and it’s running automatically. Is 1.3345v considered acceptable?
- What other settings should I tweak to aim for around 4.9 GHz?
- My cooling solution is a Deepcool 120 EX (AIO), and I’m using the Grizzly Kryonaut thermal paste.
A bit high, but I wouldn't be concerned. Those temperatures shouldn't occur during regular operation.
A constant clock speed remains unchanged, following your instructions.
VID represents the programmed factory voltage level. Core voltage should be what you observe; if the system stays on auto and uses LLC, it might be running a bit higher than ideal.
4. No clear solution, but adjusting the voltage and boosting the clock speed could help extend performance. Otherwise, you'll likely need superior cooling than a standard radiator.
4.8Ghz across 8 cores performs quite well.
Eximo, thank you for your response. The 4.8 Ghz is acceptable, but I prefer racing with VR using iRacing. Unfortunately, I notice that combining VR, Physics Calculation, and Frame preparation heavily uses CPU resources, leaving little for the GPU. The main limitation appears to be the CPU.
Thank you for your thoughts on temperature; I agree it might be a stability test rather than normal use. I appreciate your understanding. The Intel software set 105 degrees as the maximum Tj Max (my processor passed all tests, including temperature). The most demanding operation is the prime number calculation, and I doubt that running all eight cores at full capacity for such tasks will be efficient. I plan to monitor the daily usage temperatures and hope it will be satisfactory.
I recently ran a test and achieved 4.9 GHz (adjusted multiplier to 49x). The Intel Burn In Test lasted 30 minutes:
clock speed remained stable at 4.9 GHz
hottest core reached 94 degrees, with an additional 5 degrees from the 4.8 GHz
VID: 1.3793v (processor used slightly more power)
Memory frequency: 99.95
Additional questions:
1. Why does memory frequency change even when I didn’t modify any memory settings? If I enable auto-sync for all cores, does it default to the processor’s 100 MHz setting? Is this typical?
2. Should I adjust the voltage manually or try different offsets to find the optimal value? Which approach is better? (Currently, I’m using the default auto setting)
3. Are there any BIOS settings I should consider?
Apologies for writing extensively; I’ve gathered a lot of information online but haven’t found detailed personal experiences. That’s why I chose to share my own insights.
I don't have a clear idea of the frequency you're measuring. The BCLK should be around 100Mhz, and 99.95 is quite close, but the memory frequency will likely be a multiple of that. If you're using 2666Mhz memory, it would result in a 13.3 multiplier. The cache frequency is usually based on the CPU's base speed.
Auto overclocking isn't usually advised, especially when running at 2666Mhz, because motherboards often apply excessive voltage to maintain compatibility. Since you haven't checked the core voltage, I can't give a precise recommendation. You should rely more on additional monitoring tools rather than just Intel's defaults.
Offset voltages mainly affect the C-states of chips, allowing the CPU to lower its speed and use less power. They are equally risky as LLC settings and can surpass the intended core voltages if misused. Negative offsets reduce power and heat but might drop the voltage too low for proper operation, potentially causing crashes. With a fixed frequency, it's best to adjust voltages manually.
The fundamental steps in overclocking involve checking Vcore, VIN, LLC, Multiplier, and possibly cache speed. There are many more settings to fine-tune for optimal performance. There are numerous guides available; I find overclock.net helpful, though there are also board/brand-specific resources that could be more relevant.
Remember to track all voltages using a tool such as Hardware Monitor. While no single app is perfect, testing several can help you identify the most accurate readings. Sometimes they may misinterpret sensor data. Skylake, Kabylake, and Coffee Lake architectures are quite similar, so chances of finding significant issues are low.
Your performance depends largely on how well you manage a solid chip selection.
As of 12/07/2018
The percentage that can achieve an overclock at a reasonable 1.360v Vcore is around 5.1% and 5.0% respectively.
For the I7-9700K, AVX offset is set to 2.
The Intel burn test effectively simulates the most demanding tasks.
This isn’t typically what you’d do in everyday use.
The most accurate evaluation comes from your own workload.
Adjust your RAM to the optimal XMP setting first.
Track temperatures using CPU-Z and maintain the Vcore near 1.36.
CPU-Z offers a straightforward stress test for testing.
Keep an eye on temperatures with HWmonitor; expect peak temps around 100°C, with safe limits up to 85°C.
Apply speedstep and adaptive voltage to lower the multiplier and Vcore when the CPU is idle.
Just proceed with the multiplier you’re comfortable with.
A crash during a game isn’t a major issue.
If you face issues in regular use, reduce the multiplier slightly.
Eximo, thank you for your time.
1. I'm using a 100Mhz processor, which BIOS detects by default.
2. After installing CPU-Z, the readings were:
CPU Core Voltage (idle): 1.314v @ 4.9
CPU Core Voltage (during burn-in): 1.181v @ 4.9
I also checked with the AI Suite (from Asus), which confirmed the same data for Core Voltage, so I believe it's accurate.
I assumed the voltage would rise under load (burn in).
3. Why do you consider LLC to be dangerous?
My motherboard has the AVX instruction core ratio with a negative offset. Is this the same as the AVX offset?
AVX instructions use high voltage and generate heat. Fortunately, they are not that common in suage for games and such.
The offset will reduce the multiplier when avx instructions are present.
I see no harm in that at 2.
Under load, the voltage might be reduced( I think it is called vdroop)
If you get sophisticated with overclocking you can adjust those parameters.
LLC applies more voltage under load for stability.
I suppose LLC can be dangerous if it results in excessive voltage at any time.
Under load your voltages are quite good and I think you could go higher if you want.
Thank you Geofet, I wanted to clarify a few points.
The BIOS states that subtracting this value from your core ratio will give the ratio where AVG applications run. A negative offset might sometimes need a higher core voltage to keep things stable.
I’ve seen similar information before—when AVX is reduced, heat can decrease (more space for higher clock speeds), but my motherboard mentioned that negative offset could demand more voltage for stability. Is that correct?
Sorry, I don't have expertise at that level of overclocking.
The statements are a bit unclear—raising a negative offset could actually be increasing it.
You might check overclocking forums for your type of motherboard.
It might be wise to avoid adjusting that setting.