i7 4790k OC
i7 4790k OC
You're considering a 4790k and wondering if an 212 Evo is sufficient or if the H100i would be better. It's natural to worry about pump issues or leaks, but these are common concerns. Your choice depends on your comfort level with the risks versus performance gains.
I recommend using the Cryorig H7 air cooler and keeping the CPU at its standard 4.4GHz Turbo max configuration.
Although you might be able to boost it to 4.6-4.8GHz with liquid cooling or a more costly air cooler, I don’t believe the real-world benefit is significant, particularly for gaming.
I advise against using liquid coolers unless absolutely necessary. They function adequately (though some may be noisier than expected), but problems like pump failure can lead to complicated returns and reinstalls.
Air coolers tend to last without major issues, and when they do fail, it’s usually just a worn fan after several years. My advice is:
1) Cryorig H7
2) Assemble your PC
3) Apply default/XMP settings for CPU/DDR3 memory
4) set...
212 evo works fine, but you're likely to hit thermal throttle before reaching the chip's full potential. You won't need a water cooler to achieve better cooling. Consider dual tower air coolers like noctua nh-d14, Phanteks PH-TC14PE, cryorig r1, deepcool assassin, etc. Also, the motherboard you're using will impact performance.
TofuLion shares similar concerns about thermal throttling and cooling efficiency. It suggests considering dual tower air coolers instead of water coolers, mentioning models like noctua nh-d14, Phanteks PH-TC14PE, cryorig r1, and deepcool assassin. He also notes that the chosen motherboard can impact performance. He plans to use a MSI z97 gaming 5 MOBO and is evaluating more effective air coolers without a significant cost increase.
I recommend using the Cryorig H7 air cooler and keeping the CPU at its standard default, 4.4GHz Turbo max settings.
Overclocking to 4.6-4.8GHz is possible with liquid cooling or a more costly air cooler, but I don't think the benefit is significant in real-world scenarios, particularly for gaming.
I advise against using liquid coolers unless absolutely necessary. They usually work well (though some may be noisier than expected), but problems like pump failure can lead to complicated returns and reinstalls.
Air coolers tend to last longer; failures are rare and often due to worn fans after several years.
My advice is:
1) Choose the Cryorig H7
2) Assemble your PC
3) Apply default or XMP settings for CPU and DDR3 memory
4) Configure a fan profile (via motherboard software) to balance cooling and noise (target around 75°C, increasing slightly at 45°C).
If you later decide to overclock, the H7 remains affordable.
Other suggestions:
You can adjust the MULTIPLIERS to fine-tune performance without changing voltage.
a) Run the CPU and system for a few weeks to confirm stability, then
b) Increase multipliers for 3 and 4-core usage in BIOS.
This might yield results such as:
- 1-core: x44
- 2-core: x43
- 3-core: x42
- 4-core: x41
You could achieve around 4.1GHz under full load. If this happens, try setting a higher multiplier (e.g., x43 for 3/4 cores) and test with Prime95.
I understand, but I didn't want to draw attention to it. Boosting the 4790k isn't really essential and you probably won't see a significant change in most situations... it's more about passion than practicality.
I understand, but I didn't want to draw attention to it. Overclocking the 4790k isn't really essential and you won't see much change in most situations... it's more about being an enthusiast than practical. It's important to bring these points out because others might not realize them. For instance, if you have a 4.7GHz processor and adjust one to 4.3GHz, you'd see around a 9% difference in theory, but possibly a 0% to 2% edge for gaming. Video conversion could show 9%, while other tasks would notice little change. If we were discussing significant differences, it would be different, but adding a liquid cooler is tricky—it often makes things noisier and more likely to fail.
Well, certain games really gain more from faster cores. ARMA III, for example, shows a clear (even without checking FPS) variation between 4.4 GHz and 4.7 GHz. That’s an uncommon situation and most games aren’t directly influenced by CPU speed. In fact, ARMA III also displays similar differences when overclocking the GPU.
On the other side, handbrake doesn’t seem to appreciate overclocked CPUs much. There’s a distinction, but it’s not as significant as one might think. During my early attempts at stable overclocking, it even performed worse—sometimes crashing, other times running very slowly—especially when voltage was too low.
The average gamer probably doesn’t mind such a small 5 FPS gap, and some might not even notice it. But for enthusiasts like me, it can be more enjoyable to fine-tune for peak performance in a game than to actually play it.
Also, I wouldn’t use Prime95 on Haswell because certain versions could exploit the CPU excessively, causing it to overheat. I prefer Intel Extreme Tuning Utility for initial testing and OCCT for long-term stability. The best way to truly assess is by using the system itself.
I covered the overclocking possibilities. Regarding your example, increasing from 4.4GHz to 4.7GHz is below the theoretical performance limit of 7GHz.
ARMA III:
That implies the game won't boost the frame rate beyond a 7% increase (essentially hitting the CPU bottleneck). Assuming a 4% boost for ARMA III, which seems realistic under heavy CPU strain, we'd be looking at an average of 52 FPS instead of 50.
I'm not confident this would be obvious to most users.
Handbrake:
Overclocked CPUs don't necessarily perform better. The improvement is directly tied to the frequency. As I mentioned, a 10% speedup would result in a 10% faster completion (very close, but other programs may influence slightly).
Your Handbrake issues seem to stem from stability concerns with your configuration. You noted insufficient voltage, so running slower forced throttling. When setting up your CPU, always verify it performs optimally under the worst-case conditions.
Prime95 issue?
This might be true or not; I'll look into it. But based on your description of it running "hotter than expected," it likely points back to the stability problems you experienced with Handbrake due to your setup.
In short, you might want to adjust for worst-case scenarios like Prime95 to prevent future complications in demanding tasks. That's ultimately your decision.