Some overclocking questions
Some overclocking questions
I have a few questions that require clarification after upgrading my GPU. It's a GTX 1070 Asus Strix. After using Heaven Benchmark 4.0, I've recorded these results: +225 on the core clock and +500 on the memory (I might try even higher values, as others suggest).
However, while playing, I've noticed in the monitoring window that MSI Afterburner shows the core clock and memory clock fluctuating rather than staying steady. They sometimes drop to 1506mhz or stay at a constant low frequency, not reaching the maximum I set. Why is this happening? How can I ensure the GPU runs at the settings I've configured? It seems unusual that it only drops slightly, and temperatures remain within normal limits—never exceeding 61 degrees. Also, I've set the core voltage to 100% to bypass power limitations and prevent throttling.
I also observe some red irregular dots appearing on my screen. Could this be due to an excessively high core clock or memory clock? Additionally, I'm experiencing significantly lower FPS compared to others using similar components (16GB RAM at 2133mhz, i7-6700k GTX 1070, etc.) and even on a 1280x1024 monitor. Why is that the case?
Makentox :
Vlad_14 :
Ok, now that I understand why the core clock and memory clock aren't at their maximum all the time, why do I see red dots on my screen? They last for just a few milliseconds but are still noticeable. What's behind that?
memory overclock is too high
You're right? Could it also be caused by the core clock? If so, then I'll try the profiles he suggested me.
It can happen, but usually memory is the issue. It's easy to check—lower your memory settings to play the game normally, or lower your core clock for smoother performance.
Don't rely too much on stress test software; you often stay stable in testing but crash or see artifacts when playing.
I tend to overclock on...
If your card is underperforming, it’s still managing the task efficiently. You might be able to boost quality if it’s a graphics-heavy setting. I’m thinking about this because I’m seeing around 90-100-110 fps in GTA 5, which is really high for me. The game feels overwhelming, especially since tech issues are involved. Raising MSAA to 4 didn’t help much, and the same goes for other settings. This video is confusing me. Why am I getting more output at 1440p than what I see at 1280x1024? I’m also updating my monitor to a new 1080p model, so I want to check if my FPS improves there.
The described behavior can stem from a few factors. First, the card might react to load, as its speed increases only when it requires power. Second, it could throttle if temperatures go above 82°C. Third, performance and stability might be affected by the card's VRM temperatures. HWiNFO tracks VRM temps but may not function with all cards.
The behavior you mention can have two potential reasons:
1. The card "responds to load". In the game, the load isn't steady; the card only reaches its full speed when it needs power.
2. The card might slow down if temperatures go above 82C.
3. Performance and stability could also be affected by the card's VRM temperatures.
HWiNFO tracks VRM temps but doesn't work with every card.
The game I noticed this in is GTA 5, which includes indoor areas that aren't too demanding. The temperatures I mentioned are not the issue, given the card has a decent cooler (the ASUS Strix as I've said), and the case is well-ventilated with two 140mm fans pushing air in, one at the back exhausting and another on top with a watercooler. The maximum temps I've seen are around 61°C.
So it seems reasonable that the card isn't always at full capacity. Alright then. I thought the card should run at full speed all the time, but if it's not necessary, I'm not sure.
The game I saw it in was GTA 5. It includes indoor areas that don’t cause stress. As I mentioned, the temperatures aren’t an issue, given the cooler (as I’ve said before) and the case is well-ventilated with two 140mm fans pushing air in, one at the back exhausting and a water cooler on top. The maximum temperatures I’ve seen are around 61°C. So I think it’s fine that the card isn’t at its max all the time. Okay then. I thought the card should run at full capacity constantly, but if it doesn’t need that, I’m not sure. GPU temps aren’t the issue... do you know what VRM temps are? The graphics card is generally fine.
http://www.guru3d.com/articles_pages/asu...ew,10.html
Yes, that’s normal behavior. Just note: you kept it at full load all the time, which is similar to your CPU overclock and keeps your RPMs steady at 6500... that’s not ideal.
Regarding the dots, it might be screen artifacts. Try lowering the OC and see if it improves.
I’d also suggest saving several OC profiles on AB, like:
1. Stock
2. Max Stable Core
3. Max Stable Memory
4. Max Stable Combo of 2 and 3
5. Nerfed Combo for battlefield and other games.
This lets you quickly test which profile is causing the problem.
Now that I understand the reason behind the core and memory clocks not running at maximum all the time, why do I see red dots on my screen? They appear for a short time but are still visible. What could be causing this? The memory overclock setting seems too high.
Makentox :
Vlad_14 :
I understand now why the core and memory clocks aren't at their maximum all the time. Why do I see those red dots on my screen? They last for a few milliseconds but are still noticeable. What's behind that?
It seems the memory overclock is too high. Are you sure? Could it also be related to the core clock? If so, I'll try more profiles as he suggested.