Jump to content

Is there a way to bypass / override AMD's hard-cap of only +200MHz on Auto Overclock?

I have a 5600X boosting to 4850MHz at 50~ degrees, and ClockTuner by 1usmus noted it could go up to 5050MHz. While CTR overclocking is unstable at best, it did run WoW Classic at 4900GHz with CTR undervolting it to 1.375 as well, but the undervolting it does causes hard reboots in benchmarks (cinebench).

 

Since I know the CPU is capable of boosting past 4850MHz, is there any way to bypass the hard-cap AMD put up without resorting to CTR?

Link to comment
Share on other sites

Link to post
Share on other sites

No. But you're comparing apples and oranges here. For a pure overclock, you're unlikely to even do better than 200MHz. If you can get it that high. Normally you're looking at 100-150MHz.

 

What CTR does is apply an undervolt, but it does so on a curve. The problem with stability is that this is entirely a software based solution, and Ryzen is notorious for spiking voltage requirements, which it can't compensate for being just software.

 

Zen 3 now has the curve optimizer which is very similar to what CTR does, but at the hardware level. It is much more stable as a result. That is what you should be using. CTR was for Zen 2, and isn't even really supported for Zen 3.

 

The difference between an overclock and an undervolt is that with an OC you're setting a fixed clockspeed multiplier that's hopefully better than the all core boost (which is actually lower than the boost clock on the box, which is single core only), whereas an undervolt still keeps the chip's normal boosting behavior but gives it more power and thermal headroom to boost higher. For example, stock, a 5600X has a max boost clock of 4.6GHz (single core) and you might see an all core boost of 4.3GHz. With an OC, you might get the all core boost up to 4.5GHz, but you're now locked, so you will no longer see that 4.6GHz single core.

 

An undervolt, particularly via the PBO2 curve optimizer shifts the frequency/voltage curve entirely to the right, so your all core boost goes to 4.5GHz (let's say for the sake of argument), but your single core shifts up as well, to maybe 4.7 or 4.8 GHz.

 

Actual results will vary chip to chip, but Ryzen pretty much always does better overall undervolted than overclocked, simply because the boosting behavior is so good.

CPU: AMD Ryzen 9 5900X · Cooler: Artic Liquid Freezer II 280 · Motherboard: MSI MEG X570 Unify · RAM: G.skill Ripjaws V 2x16GB 3600MHz CL16 (2Rx8) · Graphics Card: ASUS GeForce RTX 3060 Ti TUF Gaming · Boot Drive: 500GB WD Black SN750 M.2 NVMe SSD · Game Drive: 2TB Crucial MX500 SATA SSD · PSU: Corsair White RM850x 850W 80+ Gold · Case: Corsair 4000D Airflow · Monitor: MSI Optix MAG342CQR 34” UWQHD 3440x1440 144Hz · Keyboard: Corsair K100 RGB Optical-Mechanical Gaming Keyboard (OPX Switch) · Mouse: Corsair Ironclaw RGB Wireless Gaming Mouse

Link to comment
Share on other sites

Link to post
Share on other sites

I'm not sure what you said tbf.

 

I know curve optimizer helps in higher end chips, but all my tests with it either proved unstable or gave no performance benefit.

 

With the auto overclock feature in PBO maxed out at +200MHz I get a stable single core boost of 4850MHz and an all-core boost of 4.6-4.65GHz in stress tests.

 

I wanted to know if it's possible to override the maximum limit of +200MHz on the Auto OC to, say +250 or +300, +350MHz, which would likely enable me to reach up to 5GHz on single core because of massive thermal headroom and a substantial power headroom compared to the hard cap of 125W on the 5600X (of which at max auto OC settings is using about 90 to 95W).

 

I had a curve settings of something about 0 0 -24 0 -24 -24 but that was also unstable and gave no better results in cinebench R23, I also have -24 0 -24 0 -24 -24 and -15 0 -15 0 -15 -15 which did not provide difference in results.

 

The 2 0 settings are for my best cores, which heavily dislike any form of negative curve optimizer.

 

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×