R5 3600 OC
R5 3600 OC
i've been using this pc for two months now, and i've seen my cpu temperatures rise quite a bit—sometimes reaching around 50°C, then jumping to about 62°C suddenly. i thought it would help to take out the 4.2ghz oc on my r5 3600 to check things out. after doing that, the temps dropped significantly, and my fps actually improved slightly when i compared 4.2ghz with 3.6ghz on ac odyssey. for example, at 4.2ghz my average fps was 69, while at 3.6ghz it was 72. this suggests something isn't quite right with the cooler or maybe the cpu itself. but it doesn’t necessarily mean your cooler is bad—just that performance improved. you might want to run more tests comparing fps under different settings to confirm.
It often lacks sufficient power, which limits its ability to reach peak performance, resulting in lower overall results. Check out a Gamers Nexus video for more details.
Usually returns drop off... @InfernalClaw After stripping your manual OC and returning to stock, it’s likely the CPU was already pushing past your manual overclock by a small margin. A 69-72 FPS average would sit right at the edge of acceptable performance. These chips excel at self-boosting and power savings. There’s little value in manually overclocking unless you’re testing specific benchmarks... which is what I do.
when i checked it, the voltage stayed the same because i didn’t know much about it. this is the voltage reading at 3.6ghz.
Observed slight adjustments in reference clocks on my rigs to fit the system. Typically sees minor shifts around 99mhz with small variations, like dropping to about 88-92mhz. Not sure about clock stretching—would love a reliable source to understand it better.
Just let the system manage the chip. You can experiment a bit here and there. SenseMi's Performance Enhancer or whatever automatic OC tools they call it—those are what I use on my ROG board. If Ryzen Master is available, you can look into the EDC and PPT limits and features of the board. I've seen that increasing the EDC helps boost speeds by 50-100 MHz depending on the load temperatures.
I'm comfortable with overclocking techniques. Spread spectrum is typically set low by default on my ROG boards, so I always verify it carefully. My goal was to lower the reference clock significantly. I might have a screenshot available for demonstration, but I don't want the discussion to stray too much from the topic. It's possible we're using different terminology here.