i5 6600k performance drops significantly after optimization.
i5 6600k performance drops significantly after optimization.
Hi everyone,
I’ve been looking into this a lot and haven’t found anything comparable to my setup. I’m reaching out for some guidance.
Current configuration:
- Motherboard: MSI z170i gaming pro
- CPU: i5 6600k
- RAM: Corsair vengeance 2x8Gb DDR4 2400MHz
- Cooler: Noctua d14
- GPU: Asus GTX970 Strix
- Power Supply: Seasonic 650 W gold
- Case: Fractal node 304
I’m experiencing unstable overclocking. Anything above the stock settings causes problems. During stress tests and benchmarks, I noticed CPU clock throttling. Fixing it involved adjusting iccmax to 256A (instead of auto) and setting turbo boost powers to 200W (again, not auto). Once that was resolved, I could run stress tests and benchmarks without throttling or errors, with all cores at full speed and 100% utilization (per hwmonitor).
However, the main concern remains:
- Benchmark scores (like Cinebench or iXTU) are significantly lower than stock performance (roughly half the score).
- Performance drops in other OS applications.
- Temperatures stay well below 60ºC.
This issue appears consistently even when I change voltage or cache clock speeds. I’ve tried toggling parameters like turbo/eist/c1e, using adaptive/override voltage, and setting CPU ratio mode to fixed/dynamic. Still no improvement.
Could anyone point out what I might be missing?
Thanks in advance for your advice.
Best,
José
Which stress tests are being applied and for how duration? At 4.4 GHz, what was the maximum vcore attempted? You mentioned concerns about performance beyond the standard settings. Which specific frequency and vcore pairs were tested?
Hi ihog,
I began at 4.3 GHz with 1.24 V (lowest voltage without errors on prime 95 small fft for 20 min) up to 4.6 GHz at 1.39 GHz over 20 minutes. One of the interim settings was 4.4 at 1.28, and I’m targeting this speed for continuous use.
The tests were primarily prime 95 for 20 minutes, with benchmarks like xtu and cinebench. Everything finished without errors, running at full capacity, no throttling, no overheating, and 100% utilization, yet the actual performance is lower than expected. I don’t understand why!