New to Ryzen and overclocking. Please advise.
New to Ryzen and overclocking. Please advise.
I understand, this is the reason I generally suggest users either avoid installing (completely) or at least temporarily disable any bundled motherboard software or utilities from their manufacturer. Any untested or known problematic overlays or monitoring tools like CAM, various mods and plug-ins, etc., should be removed. Even better, when encountering an issue with no apparent solution, performing a clean Windows installation is often the most effective starting point to determine if problems exist in a clean environment.
It didn't appear to be a clear concern for your situation, which is why I didn't advise it. However, doing so can be helpful if you reach a dead end, regardless of the issue's nature.
The smallest FFT isn’t something we consider a test metric. What we mean by Small FFT is what we actually use. All your thermal compliance and stability testing procedures, which are quite different, can be found towards the end of my guide here: CPU overclocking guide and tutorial for beginners. Covering the basics—if you’re here, it’s likely you’re seeking help with basic overclocking. Or maybe you’re a seasoned overclocker curious if I’ve missed any insights you might have. Rest assured, I don’t. This is only intended as...
The big question here is, how much better is this OC of yours (Temp, CB R20 results, voltage, etc.) compared to run the Ryzen 5 3600 with PB enable and PBO and AutoOC disable?
With those settings I wrote up there I get a CB R20 multicore score of 3632 and a single core score of 482, So Im not sure you are getting a lot of extra performance out f it. (I get this results with the system you can see in my Sig)
Heck, in fact you could just try with PBO enable too, after all you have a 280mm aio water cooler
Cheers!
In stock configurations, Prime95 wouldn't run at 4.2. It would mostly stay around 3.8ghz with occasional hits near 4.0ghz. Temperatures would sit in the upper 70s. This is quite similar to the current setup. I wasn’t reaching the boost threshold of 4.2, and I understand it’s not designed for sustained high loads. That’s why my 'OC' is just pushing it to the boost speeds without going over 4.2ghz.
In Cinebench, my multi-score improved from 3116 to 3676. The time measurement has increased from 8904 to 9086. The biggest gain came from better handling of the boost behavior that previously caused constant fan acceleration (likely due to excessive voltage under PBO). Now it stays within a fixed voltage of 1.125v, even with fan smoothing and reduced responsiveness. I only reacted when temperatures were high, which was unexpected for me, resulting in very high temps during regular gaming. Now I’m not seeing the fan ramping, and I haven’t tested gaming yet.
Thank you for your feedback. Is a constant 1.125v really harmful? Could it accelerate wear faster? And is there a way to adjust this fixed voltage so it drops when not under load?
Thank you for the guides. I'll dedicate some time to review the Ryzen Master parts. Cheers.
Yes, there is. In the BIOS, adjust all the listed settings as shown here:
Cool N Quiet - Enabled
Core CPPC - Enabled
CPPC preferred cores - Enabled
Advanced/Global C-states - Enabled
In Windows, go to power options, choose a power plan such as Ryzen balanced or balanced, then proceed to advanced settings for that plan and configure the minimum processor power state to 5% and maximum to 100%, ensuring it matches your current setting.
It's great to hear that. I'll search for those BIOS settings in BIOS, some familiar and others not. AsRock BIOS doesn't provide much detail. I plan to adjust the cores and voltages in BIOS as well, if anyone knows the specific words used by BIOS compared to Ryzen Master. I was using RM mainly for quick stability checks. If Prime95 can endure an hour of small FFT tests, I'll proceed with BIOS adjustments and then run an overnight test in Prime95.
Power plan in Windows is configured as Ryzen balanced, with the minimum state at 5% and maximum at 100%.
Nothing I recommended matches a configuration found in Ryzen master, so it’s unclear why you mentioned that?
All settings are present in the BIOS and within Windows configurations or control panels.
If you run Prime95, make sure to select "Small FFT" instead of "Smallest FFT." Also, turn off all AVX options. Certain choices might not be available until other AVX settings are disabled. These AVX disables appear in the main torture test window.
You don’t need to let it run for an hour just to check thermal compliance. A 15-minute test is sufficient; if it stays below 80°C during that time, it’s likely safe regardless of duration. That’s the main useful measure for Small FFT in our context.
Once you confirm your overclock is fully thermally stable and ready for stability checks, this differs from thermal compliance testing. Thermal compliance should use a full-load steady-state tool operating near 100% TDP, whereas stability testing typically employs variable load or workload utilities across the board. Realbench’s stress test with half the installed memory is the first I’d perform. Initial trials last around an hour, followed by an 8-hour stability verification.
After that, other stability signs include executing four passes of Memtest86, applying special Prime95 configurations, and mixing various OCCT test combinations.
I was just confirming that I haven't altered anything in BIOS beyond adjusting fans and setting XMP 3600 with fclk at 1800. I ran memtest86 on the memory in this configuration and all four passes came back clean. I used RM to experiment with different settings until I felt confident enough to tweak BIOS further. Good suggestions. After making BIOS changes, I plan to perform thermal tests with Prime95 and stability checks using Realbench. I'm aware that overclocking the CPU can impact RAM stability, so I'll run another memtest86 once the CPU is stable. Thanks for your assistance. I'll follow up in a few hours.