Stabilizing and locking core clock on a Pascal GPU? - SAFE OR NOT?

May 14, 2018
26
0
40
Hey guys, I came across this video where you can lock your gpu core clock to a certain amount, which will give you the maximum performance for your gpu. Is this safe or not?

Note: This only works on pascal cards as they tend to overclock themselves, so when you lock it to the highest amount, the gpu will consistently run on that specific core clock.

Video: https://www.youtube.com/watch?v=LLMe2iS_IIU
 
Yeah but sometimes the card occasionally drops its own core clock which may cause micro-stuttering and unnecessary fps drops but locking it to the highest v/frequency eliminates these. So I'm asking if it's safe or not because it offers more performance and eliminate annoying fps drops.
 
I haven't watched the video.

When it drops, it's because the temps are too high. It's called thermal throttling. The GPU is trying to save itself. Locking the cores at a certain frequency and not allowing it to go lower can kill the card. FPS drops are usually caused by the CPU having an issue. A GPU should be able to put out the same FPS all the time. But if there is a delay in getting info from the drive, or if the CPU is thermal throttling, there can be a drop in FPS. Locking a GPU at an overclocked frequency is bad. Locking it at stock clocks is a waste of time. As mentioned above, just let the card worry about it's own frequency.
 
Can you be more specific? I'm sorry, I'm quite new to this. But I'm sure my temps are fine. My gpu is the zotac gtx 1050 ti oc which has dual fans and it stays 60° to 65° when playing games and I have it locked at 1949 mhz / 1049 mv using msi afterburner. And my processor stays around the same temps. 60° to 63° when playing games.
 
I didn't notice the video mentioning anything about micro-stuttering, though I could've missed it.

There's no reason for the GPU to drop its clocks under load. Either the load has been reduced, or it is thermal-throttling. Why else would the clocks drop?

Forcing it to stay at maximum at all times does not sound safe to me.
 
Lets say you have a badly fragmented 5400RPM "green" drive. If you try to load more data from the drive and it has to spin back up, find these random bits of data spread out all over the drive, and then send them to the CPU to get worked on, I can totally see you having skips or pauses in your frame rates. It has NOTHING to do with the GPU dropping it's frequency or high temps. It has everything to do with a slow drive slowing everything down. Or "bottlenecking" the system. You really need to look at your system as a whole to see what's going on. You can't just look at your GPU dropping the frequency and say it's the GPUs problem. Why is it dropping? Once you know why, then you can fix it. The fix is probably not "lock the frequency". Nvidia is smarter than that.
 
@King_V, I saw a guy in the comments who said it eliminated the micro stutters and it truly does, I had random fps drops here and there in pubg but locking the gpu to its highest frequency did it for me. But I want to make sure its safe first. My temps are fine even though its overclocked and locked at 1936 mhz, its always around 60c so I don't see a problem tbh
 
@4745454b, Yeah you're right man you made a good point. How do I figure out if that's the problem though? I only use MSI Afterburner. My temperatures are:

GPU:
Idle: 40c to 44c (locked at 1936mhz)
Heavy load: 60c to 67c (locked at 1936mhz)

CPU:
Idle: I'm not sure as afterburner merely shows proc temps when I'm gaming.
Heavy load: 60c to 64c.
 
System specs? Not all CPUs have the same thermal throttle point. I can't check your drive or fragmentation level either. Play games. Do the frame rate drops happen at the same time/points? Is it loading data?

As for temps, try something like this.

https://www.hwinfo.com/

Load it, game. Go back and look at the max temps. That's how the other guy found out about his VRMs.