Inquiry Regarding My CPU Performance: Intel Core i7-7700K
Inquiry Regarding My CPU Performance: Intel Core i7-7700K
Hey there.
I'm running an Intel Core i7-7700K overclocked to 4.8GHz on my MSI Z270 SLI Plus motherboard with the OC Genie software.
There are a couple of things I’m wondering.
First, am I actually noticing any improvement in gaming performance, or am I just perceiving it? I play a range of titles—from Fallout 4 and Skyrim (both heavily modified) to Ghost Recon: Wildlands and Battlefield 4.
Second, my temperature readings are mixed. Corsair Link says 82°C, while Core Temp shows 73°C. If the CPU is really reaching 82°C, why? And if it’s not getting hot enough, how do I fix this?
Here’s a summary of my build:
- Case: Corsair Crystal Series 460X RGB Mid-Tower
- CPU: Intel Core i7-7700K OC'd to 4.8GHz
- CPU cooler: Corsair H100i V2 with fans at max (2,700 RPM)
- GPU: EVGA GTX 1070 SC
- RAM: Corsair Vengeance RGB DDR4 3,000MHz OC'd to 3,200MHz
- PSU: EVGA 850W Bronze
- Thermal paste: Arctic Silver 5, horizontal line method
- Airflow: 3x 120mm front intake, 1x 120mm rear exhaust, 2x 120mm radiator fans (top exhaust)
I’m concerned about potential CPU damage from overclocking and not seeing any real performance gains. Yet, I’ve heard that open-world games like Skyrim, Fallout 4, and Wildlands can actually benefit from an OC. At the same time, I don’t understand why my CPU would be getting so hot—doesn’t it have enough cooling?
Grr. I really don’t know what to think about here!
I truly hope CPU makers would either be more consistent with silicon or switch to a completely different material. Isn't that what people are doing when they replace the delid with liquid metal? They're swapping the silicon.
silicon is simply what the CPU is made of.
what you're referring to is TIM – thermal interface material positioned between the CPU die (which is silicon) and the IHS (integrated heat spreader), that metal component you see.
http://www.overclock.net/content/type/61...0/flags/LL
the answer to the question in the picture is YES, and on the right side is the backside of the IHS.
Some brands use auto overclocking as a convenient method, assuming it's a secure option. However, this can lead to unexpected voltage spikes if you manually adjust the settings in BIOS. My 7700K was set to 5.2ghz via manual BIOS tuning, yet I have superior cooling, but even with that, manual overclocking can push it to 5ghz. My peak core stress at 5.2ghz during Intel Burn Test reached 74°C, gaming cores hit 53°C, and the best gaming performance was 57°C while running FarCry 4. The full guide provides more details.
auto overclocking is problematic because it forces significantly more voltage than necessary, which manufacturers aim to handle so even the weakest CPU can run smoothly without failing. Many CPUs can achieve those speeds at much lower voltages, but this leads to excessively high temperatures.
Additionally, Intel's 7th generation models are often associated with thermal challenges because of the thermal interface material used under the IHS.
You might consider these options:
1. Manually overclock your CPU, keeping voltage below 1.3V (even that is excessive).
2. Perform a CPU defrost.
Someone is suggesting that using auto overclocking on any motherboard brand is a shortcut, but they warn that it can lead to the CPU voltage rising beyond safe limits if you were doing it in BIOS. The person shares their experience with a 7700K CPU, which was manually overclocked to 5.2ghz and notes they have better cooling than others, yet still manage to push it to 5ghz with full control. They also mention temperature readings during intensive tasks: 74°C for the highest core at 5.2ghz, 53°C during gaming, and 57°C while running FarCry 4 and playing hard. The link provided discusses overclocking statistics. The reason given for choosing this method was advice from a friend about MSI's support for their OC Genie, implying that if the CPU is damaged, MSI will cover the cost. The author admits to not knowing much about manual overclocking but remains open to the possibility it could result in lower temperatures compared to 4.8GHz.
Here’s the updated version:
Last night after returning from work, I decided to overclock my CPU manually rather than letting the BIOS handle it automatically. By maintaining a stable 4.8GHz, I was able to reduce the voltage considerably. The OC Genie was delivering nearly 1.4 volts, which likely caused the high temperatures. After several adjustments and tests, I found a consistent setting of 1.15 volts. During testing, Prime95 ran for more than an hour at this voltage, reaching a peak temperature of 77°C. Playing GTA V on Ultra for 30 minutes kept me around 55°C, while Ghost Recon: Wildlands on Ultra for the same duration reached about 61°C. These results represent a significant improvement over previous readings, and everything appears stable so far. It looks like I’ve reached a good point at 1.15 volts.
Good job.
With such low voltage, it seems you could go beyond 5Ghz.
Using prime95 in various modes confirms near 100% stability, but it doesn't match the typical load patterns we encounter regularly. For instance, even at full CPU usage during gaming, the CPU isn't significantly stressed by prime95.
Tasks like cinabench and other benchmark tools are better suited for evaluating stability in everyday overclocked setups.
Good job. With such low voltage, it seems possible to reach 5Ghz. Running CPU stress tests with prime95 confirms near-perfect stability, but it doesn't match the performance of typical daily tasks. Benchmarks and other tools are better suited for evaluating stability in regular use. I'm not sure if I should go beyond 4.8Ghz either. Even at 5.0GHz, the difference might be negligible, and raising the voltage could counteract my goal of minimizing temperatures. What testing approach would you recommend?
You're correct, the differences won't be obvious. It's enjoyable to observe how much you can extend the overclock.
Several factors beyond core voltage influence stability. Adjustments include cache multiplier, cache voltage, load line calibration, and turning off unused features such as virtualization support or power-saving options. Even for seasoned overclockers, changing VCCIN or VRM switching frequency isn't always clear.
You're already on the right track with stability checks. A quick test with prime95 is recommended, but be aware that high power consumption could stress components like the VRM on the motherboard. If issues arise, the system may crash or become unresponsive within minutes. In a stable condition, prime95 will display errors. Cinabench15 serves as a good CPU benchmark, and RealBench in benchmark mode is also useful.
Additionally, normal operation and monitoring error codes from Windows when crashes occur are important to watch for.
You're correct, you won't easily notice the changes.
It's just enjoyable to observe how much you can extend the overclock.
Beyond core voltage, numerous factors influence stability. There are adjustments on the more apparent side, such as modifying the cache multiplier, adjusting cache voltage, fine-tuning the load line calibration, or turning off unnecessary features like virtualization support or certain power-saving options. Even for seasoned overclockers, some aspects remain unclear, like enabling VCCIN or VRM switching frequency.
You're already on the right track with stability checks. A quick test with Prime95 is recommended, but keep in mind that the high power consumption could stress components such as the VRM on the motherboard. If issues arise, the system may crash or become unresponsive within a few minutes. In most stable scenarios, Prime95 will display errors. Cinabench15 serves as a good benchmark for CPU performance. RealBench in benchmark mode is also useful.
Additionally, normal usage should be considered, and if crashes occur, paying attention to the error codes provided by Windows is wise.
It makes sense why many people become deeply invested in overclocking... It's been a surprisingly rewarding process of discovering exactly what I can achieve.
In my situation, however, my goal is pure performance. Since the difference between 4.8GHz and 5.0GHz is minimal beyond raising voltage—which again raises temperatures—I'll likely stick at 4.8GHz for the near future.
My motivation to identify the lower CPU limit comes from wanting the lowest possible temperatures while still achieving that 4.8GHz overclock. I'm new to this area, so I lack much knowledge about optimal stress-testing programs or the best methods to verify stability. In short, I understand that excessive voltage or insufficient voltage can trigger a Blue Screen of Death, freezing, crashes, or boot failures.
I've been playing Ghost Recon: Wildlands for about an hour at 1.155 volts, and it appears to be functioning well, though results depend on time.
Regarding other adjustments during overclocking, I'm not sure what changes to make or how they affect performance. The only modifications I've made were setting the CPU Ratio to 48 for 4.8GHz, disabling Intel Turbo Boost, switching the Ratio Mode to Fixed Mode, and fine-tuning the Core Voltage. All other settings remained unchanged.