Testing the voltage boundaries while pushing my GTX 1050 TI to its highest setting
Testing the voltage boundaries while pushing my GTX 1050 TI to its highest setting
Hello, I'm attempting to push my 1050 Ti to its maximum potential. The core clock is currently at +155MHz and memory clock at +925MHz, which are stable enough for some benchmarks with temperatures around 59°C. Although the core voltage is set to full power, it reaches its limit during high usage. Lowering the clock frequency stops hitting that limit. Should I stick with the lower settings or can I increase the voltage? Also, I believe the Superposition GPU performance test isn't sufficient for stability checks—would you suggest a free software for that?
When testing an overclock, it's important to observe which factor provides better gains—whether it's a higher core clock or faster memory. I would begin by resetting and checking how much the core clock can be increased before instability occurs, then return to the last stable speed. Next, focus on adjusting the memory clock and determine its maximum stable range. Also, keep in mind that voltage is primarily limited by power constraints at 100 percent.
Superposition should be fine i would push the core clock dont touch the memory till you can push the clock speed. once you have a core clock thats the max you can go without it crashing.
then make note of that core clock speed and set it to that number.
then start moving the memory clock in small numbers of 25 to 50 then 75 then 100 once you hit a unstable memory go back to previous clock e.g you hit 100 from 75 so you go back to 75 and try 80 to 85.
then that gives you your max core clock with your max memory clock save it as a profile.
Consider that everything you enter into these apps is merely guidance for the card. The card's firmware includes safeguards that could stop it from performing exactly what you intended if it determines it isn't safe.
Regarding stability, ask yourself your goals. If it's simply about avoiding crashes, Superposition works. However, if you need consistent performance metrics like FPS over a specific time frame, tools such as 3DMark can help without requiring manual data collection.
I managed to get it working on superposition with the maximum clock, but when opening games or using userbenchmark the GPU power limit is reached and the app crashes. I adjusted the power settings for dynamic clock adjustment and userbenchmark didn't crash. The peak power usage was 95.6W with a TDP of 75W, which is at 125% of the maximum during peak. Should I lower the clock speeds to prevent hitting the power limit? Also, do voltage or clock speeds influence power use?
Voltage and clock speed significantly influence power usage. This is why makers keep hardware close to its original design—higher speeds increase power consumption, reduce efficiency, generate more heat, and accelerate component wear over time. These are all compromises made to gain extra performance.
The biggest obstacle is the built-in Gpu Boost feature. It controls the GPU excessively, often changing settings without your input if it doesn’t match the card’s specifications.
When trying to overclock, it’s important to keep custom configurations stable so you can evaluate their reliability.
Gpu Boost interferes with that process.
Frequent hits against power limits cause it to drop to a lower performance curve.
It also lacks a definitive GPU benchmark; this is why results vary across applications and can sometimes fail unexpectedly.
You won’t reach the true peak because the card constantly adjusts.