Can I make my i9-9900k run faster than its normal speed?
Can I make my i9-9900k run faster than its normal speed?
It doesn't make sense just because of how things are set up. A system like yours isn't necessarily better for everyone else either. While this setup works fine, if you want your computer to get cooler than usual, putting a radiator on the front helps much more since it pulls cool outside air in and lets hot inside air out. The gain is small but real.
Well technically if those cores are running at 4.7GHz that is an overclock or turbo boost because the normal speed without boosting is usually around 3.8GHz, I think. The main question here is whether you can push it to 5GHz or not and that depends on how good your cooling is. First off, running Prime95 gives a 10 to 12-degree Celsius higher temperature compared to regular tests like Aida64 or other stress tests. So if you want to check things under full load, use a free benchmark app and those stressed cores will really be around 68 degrees C. You can bet the chip will go up to 90 degrees C and still work fine, so yes, you can try pushing for an overclock of more than 4.7GHz on all cores or something like that. 🤷♀️😲👶☮👩🦲
rickypicky5 , Even though your test setup is okay and you have a good front-mounted radiator, you are still short on cooling power if you want to push your 9900K hard. A standard 240mm fan isn't enough for that; you need something bigger like an extra 80mm or a custom loop instead. Turtle Rig , Just like most other people do here, you seem to think some things are true when they aren't. We have to clear up these wrong ideas so the original poster can learn what's real and keep our other forum members happy too. "Firstly" ... Running Prime95 Small FFTs without AVX is actually considered "full load". When we look at all the different factors that affect heat, like your room temperature or just how well your computer works, I wouldn't guess that your 9900K would get up to 68 degrees Celsius. These numbers are thrown around everywhere online, and they aren't always true. Since you didn't tell me what it's really like for you, we can't say exactly how hot the CPU gets with just a small amount of info. "Full load" is a fancy word that means different things to different people, so using it without knowing your specific situation is risky. Games, apps, streaming, making videos, and most programs only do part of their work at one time, causing the CPU temperature to go up and down in ways we don't know well enough to judge performance by them. Stress tests split into two types: some keep the processor busy but switch back and forth (like fluctuating loads), while others run it steady all the way through (like thermal stress). Prime95 Small FFTs without AVX is perfect for checking how heat builds up because it runs at 100% capacity with steady temperatures. No other free tool can copy Intel's exact test pattern so well. If you use a program that doesn't push or pull on your processor enough, you might get a wrong picture of real performance. Here is a breakdown of tools compared by how much power they ask for: Even though these tests go up to 130% of the CPU's normal power limit, Windows Task Manager counts every test as if it used all your CPU power at once. The actual heat comes from how much electricity you use (measured in Watts), not just how busy the processor feels. Prime95 runs a steady load even when you overclock past its limits. If your temperature stays under 80 degrees Celsius, your CPU can handle tough real-world jobs without breaking down. You can see that Prime95 with AVX adds a huge jump to the power usage (about 130%). But Intel says their official specs are based on tests without AVX, so we run our Prime95 test that way too. If you use OCCT and pick its first simple test mode with no special math tricks enabled, it runs at about 97% capacity, which is almost the same steady heat as our Prime95 test. The main issue with AIDA64 is it has so many options for stress testing (CPU, FPU, Cache, Memory) that this creates too many different scenarios to check properly. Some of those tests are way over the power limit like 115%, while others hover around 70% or just barely pass the limit. All the other choices on AIDA64 are good for checking if things stay stable during short bursts, but they aren't good enough to judge long-term heat issues. Here is a quick list of what we compare: small FFTs, blends, linpack, and IntelBurn Test. The steady pattern from Small FFTs lets us measure core temperature accurately so the CPU, cooler, socket, motherboard, and power supply can cool down properly without stress. Even though Intel says the 9900K's "Throttle" limit is 100 degrees Celsius, experienced builders and experts agree that you should stay a little below that mark to avoid overheating or failing later. So here are the safe temperature zones: go above 85 degrees if you want to be sure, but stay under 80 if possible. Both temps depend on how hot your room gets because heat moves better in warm rooms. If you'd like more details on this, check out our Intel CPU Temperature Guide from 2020. This page is always here for everyone's benefit.