Question Strange real-life outcomes following a mild OC on FX-4100
Question Strange real-life outcomes following a mild OC on FX-4100
The kid is interested in playing Apex Legends and needs an FX-4350 @ 4.2. I received a free FX-4100 which works but causes lower FPS. Using old BIOS settings, I found some online tutorials, adjusted all configurations as recommended, increased voltage by .050V, and raised the chip clock from 3.6 to 4.0. I tested Realbench, achieving 4028mhz at 54°C and 1.376V. It seems sufficient for now, but I’d prefer a bit more cooling before targeting 4.2.
Last night, I ran a CPUID test while the kid played the game. The package temperature was 57°C, and core temperatures reached 90°C. That’s 90C—can that really happen? Voltage stabilized at 1.472V on the VCORE (the one I’m modifying), which is acceptable, and CPU clocks were close to 4028 but inconsistent.
This is my first attempt at overclocking an AMD chip. Before this, I only succeeded with a 2500k (success) and a 4790k (no need for OC), and I only managed .2 before voltage spiked too high, so I gave up.
What am I missing here?
I just repeated the test, cores reached 225C which is about 450 degrees F, yet nothing ignited, which is unusual. The package temperature stayed at 55 today. All four cores hit 4235mhz but remain around 4.0. Voltage stayed steady at 1.376. The chip is delivering 60 watts each time it runs.
I have a larger cooler coming, I'll attach it and check out the outcome.
CPUID is providing incorrect readings, especially on AMD platforms, and the temperature limits for FX processors are 90°C package and 60°C core. Consider using AMD Overdrive instead.
Got some varied outcomes after a few Apex sessions. All four cores hit 47.5C, max voltage reached 1.38V, average around 1.366V. The zero, one, and two temps seem to peak at 36, 54, and 58 degrees. With CPUID I saw one fan spinning at 625,000 RPM now at 3557, which feels pretty fast but closer to what it should be. If these readings are accurate, spending 22 bucks on an aftermarket cooler might not have been worth it. Would adding another 0.025V be safe with the extra 0.05V already in place?
Another odd occurrence. The chip's temperature never drops, it keeps rising. The 47.5 degrees happened after about 10 minutes on Apex. Fifteen minutes on realbench pushed it up to 48 degrees from 17. Once I turned off the real bench, the chip kept climbing (showing a 2% usage rate) and hit 50 degrees after just 10 minutes of idling. It's simply heating up more. The fan is spinning, and the heat sink feels quite warm.
Did you take advantage of AMD Overdrive? It reduces core temperature by lowering the reading from 60°C down to a range between the minimum and maximum core temperatures.
If you think the fan is the issue, look into your BIOS settings—check if the temperature-to-RPM curve can be adjusted. It should be set from a minimum of 30°C up to about 50-55°C.
You might also try setting CPU_Fan to its highest speed; it will be very loud but will help you understand the situation.
By the way, is this an OEM cooler?
Wait. Instead of displaying the CPU temperature, it shows the maximum allowable CPU temperature minus the actual temperature to indicate how much room you have before things get problematic? So when I see it idling at 60, it's really showing that it's around 11? Well, 11 can't be correct—it's 24 in my house, but is that really happening here? That's the most confusing thing I've heard all week.
I installed my Gammaxx cooler last night and adjusted the settings accordingly. I set the clock to 4.3, kept the voltage at .050 high, and let the game handle video quality, which turned out too high. I adjusted shadows and highlights to keep 60-70 fps during intense moments, and noticed a 30-degree thermal margin with normal fan speeds and a slight increase in voltage (around .005). The maximum reached was 1.38. I plan to raise it to 4.5 if everything works well. Thanks for your assistance.