Is there anyone utilizing the Nvidia control panel for optimal performance?
Is there anyone utilizing the Nvidia control panel for optimal performance?
Hello, I understand this might have been previously asked but isn't found. Are you asking if the Nvidia control panel uses "prefer maximum performance" during overclocking? I'm experiencing reduced scores when it's set to that option. My setup is an Asus TUF RTX 3080 Ti with power limits at max, where core temperatures cap at 64°C and memory at 80°C. Could you let me know what you're using or share your current settings?
Thank you for your feedback. I've reviewed many overclocking tutorials and noticed they often aim for peak performance. However, my GPU speeds vary regardless of the settings, so it's unclear if any configuration truly enhances efficiency.
During fluctuations in timekeeping is crucial. It's typical for GPUs to reduce their clock speed to hundreds of MHz when idle (such as dropping to around 300 MHz) due to power conservation. The timing of this benefit is detailed in the provided documentation link.
You could consider downclocking your GPU, even though it seems counterintuitive. My scores in superposition were significantly higher—200 to 300 points—when I did that. Yes, I always set the preference for maximum performance on every test.
Overclocking instructions are ineffective. They are tailored to particular situations rather than being universally applicable.
If you boost your card's power to 107% and play games like Amazon's New World, your card may overheat if it doesn't instantly reduce power as limits exceed 120%. Even a 100% limit results in performance averages around 107%-109%. This extreme scenario isn't unique but illustrates the risk.
Increasing maximum power limits raises voltage and current (power), which can harm GPU boost by causing temperature throttling.
I recommend using TimeSpy to monitor the card, beginning with the highest power setting, then gradually lowering it while observing results. If performance stays stable or improves, reduce the limit until you notice a decline in scores. If the score drops immediately, the issue likely lies elsewhere—such as GPU temperature, airflow, memory speeds, or GPU voltage—not just the power limit.
These guides serve as helpful references but aren't absolute rules and may not apply equally to every card.
Thank you for your clarification. I understand the situation better now. During intense usage like in Port Royal, when GPU utilization reaches maximum and temperatures remain stable, it’s puzzling why the card doesn’t reach higher boost frequencies. Given that power limits and temps are acceptable, I’m wondering if there’s an unnoticed factor or limitation affecting the boost performance.
Many individuals are using the card, yet I believe I'm not maximizing its full capabilities, particularly when my temperatures are normal.
Thank you for your response. I haven't come across dropped power limits before, but I'm willing to experiment and see if it affects my results. I'm interested in maximizing my card's performance, particularly since my temperatures are currently very low.
The 3090 I had was reaching 110°C during a run on the Heaven benchmark, so I decided to return it. This card typically reaches around 82°C, which gives me some flexibility to push it a bit higher.
If I maintain my current settings—overclock at +125 cores and memory at +750—I think adjusting the power limit down slightly by moving the slider could trigger a noticeable change in my scores. Thanks for the helpful information.
When you clock your card, do you input something like a -250 on the core, adjust it, use the curve editor, and then set values such as 850mv while changing the clocks to something like 1800mhz? I've seen conflicting advice about whether applying a -250 to the core is correct for undervolting. I'm relying on "expert" YouTube guides for this.