Optimize your i5-12600K by increasing its clock speed.
Optimize your i5-12600K by increasing its clock speed.
You're trying to adjust your system's voltage settings in XTU. It looks like the UEFI version lacks a specific undervolt protection toggle you were considering. You're currently using PL1 at 155W and finding it manageable without overheating, but you're unsure if lowering it further is safe. Intel's Speed Optimizer raises PL1 significantly, though it doesn't fine-tune the multipliers much. The advice from various sources seems inconsistent. You haven't yet triggered a blue screen, but you want to avoid risky changes. To help others, focus on checking your CPU cooler specifications and overall thermal performance before making adjustments.
When pushing the system to its limits, aim for maximum power usage. This isn’t about running without any restrictions during normal operation, but rather ensuring the power caps don’t interfere with stress testing. It helps avoid unexpected crashes after high-load tests where the system struggles to respond quickly enough. For performance checks or stress tests, stick to less demanding tools—stress tests on 2024 are generally lighter than typical benchmarks. If boosting performance is the goal, prioritize the P cores first; aim for around 50x speedup on a 12,600K chip, which most modern chips can reach (51x usually possible with good cooling). Reserve E cores for later once the P cores are stable. After that, experiment with E cores and finally adjust the ring clock. If you don’t need the extra multicore power from E cores, turn them off early—they boost clock speeds but add complexity.
It wouldn't exceed 200 watts, though I'm not certain about the exact limits. It seems Intel chose this setting to prevent overclocking automatically. That approach felt unusual at first, but it's new territory for me. I confirmed that PL1 and PL2 can be unlimited, but that doesn’t seem right. I’m unsure what all the numbers mean. The internal stress test is less intense than Cinebench, which makes sense. In real-world use, I doubt I could reach those levels. After testing, CPU temperatures fell into the 30s, and I was pushing 50x across cores—Intel usually reduces performance on multiple cores. It probably has some automatic voltage control that changes a lot. There might have been a brief spike near 1.4 volts, but not sustained. I didn’t adjust any voltage offsets myself. This is the first CPU with more than two cores I own—I’m wondering if I really need them.
Maxed out reaches 4095W, but real performance stays close to 200W under full load. This setup avoids power limits affecting clock speed, allowing accurate frequency testing. It's meant for short bursts during overclocking sessions, not continuous use. After stress tests, you can return to normal settings to manage heat. Running at these extremes isn't risky if temperatures stay stable. The Y-Cruncher is recommended for reliable stress checks. If you're using the system for gaming, consider limiting cores and disabling E cores for better stability.
Was intended to play games (playing isn't frequent), also create digital art, edit videos, and stream light content. Interested in trying VR, though I'm unsure if it's feasible. Don't need top resolution or fast refresh rates. With two functional PCs, I could potentially capture footage on one and stream it to the other if needed.
For those tasks, you likely don't require the main components, since light video editing only needs features like stream highlights. Turning them off will boost your clock speed, improving gaming performance. When streaming, use the iGPU for encoding (QuickSync) instead of a separate machine.
Thanks for bringing this up. I realized I might have missed something because the other system uses a custom motherboard and UEFI, which prevents it from working with that setup. Appreciate your advice on stress testing. The results indicate either a major installation error or insufficient cooling. I’m feeling frustrated. I enjoy tinkering with PCs, but I also appreciate well-built machines with proper components. Without power limits, the system struggles with 49x on the P-cores; I was watching and stopped it when temperatures reached 95°C, which happens rapidly. It seems it can’t handle anything beyond its stock specs (45x if all cores are active). I thought there would be some flexibility, given the 350w TDP rating for an AIO. However, the 12600k model typically runs between 230–235w with 49x P-cores removed, which feels high considering the 125w TDP and a max rating around 150w. I’m also puzzled by the Intel stock profile in XTU: it shows 241w PL2 56 seconds at 45x all cores, but the case doesn’t fit an air cooler due to space constraints. I suspect a VRM fan would be necessary, as recommended by Asrock. I assumed they tested with a downwind cooler, but without access to force airflow onto the socket (unless the AIO has one), it’s hard to replicate. This feels partly my mistake—I really wanted to try that build. It would help if there were clearer guidelines on cooling performance across different products.
Maybe I'm not sure about what I'm doing. I hope I didn't make a mistake. Windows seems to dislike Y-cruncher, and I'm not sure how to use it (the instructions don’t make sense). I tried Prime95 small FFT with AVX turned off, even though I had it enabled before. I know the CPU performs better with AVX, but some people disable it. I’m unsure if that’s a good approach. It significantly reduced power consumption by turning AVX off, yet I ran tests at normal speeds. There are so many conflicting opinions, it’s hard to decide what’s true. I set the clock to 49x because that’s the standard boost setting for single-core, but is this really an overclock for multiple cores on the P cores?