Simple ways to increase GPU performance for FREE

Summary of the technological concepts + “free” GPU performance steps

The speaker presents a tutorial focused on increasing FPS from an already-installed GPU at no cost. The improvements are most relevant when the system is GPU-bound—meaning the GPU is the limiting factor, not the CPU. He uses a controlled testing approach and tracks how common background factors and driver/control-panel settings affect 3DMark Time Spy results.

---

Testing methodology / validation approach

• Uses a “dyno” style benchmarking mindset: treat GPU benchmarks like measurable track tests.
• Prefers synthetic benchmarks—specifically 3DMark Time Spy—over gameplay because they’re more consistent for comparing changes.
• Measures:
  • Graphics score
  • Average FPS
• Creates a baseline run, then repeats the benchmark after each change.

---

Test system + bottleneck concept

• Example setup mentioned:
  • Intel i7-8700K
  • 16GB RAM
• The goal is to deliberately operate in a condition where performance is GPU-bound, avoiding CPU-bottleneck confounding (framed as “a different kind of bottleneck” emphasis).
• GPU referenced for the example:
  • RTX 2060 (not “2060 Super”)

---

“Free” background/process checks before tuning

The speaker checks whether closing common apps actually matters:

• Mentions Chrome with many tabs and multiple game launchers open.
• Notes game launchers may run background activities (e.g., scrolling, ads/demos) that could affect performance via background CPU/GPU usage.
• Also discusses launcher download behavior:
  • Steam downloads pause Steam, but EA / Battle.net / Epic may not pause in the same way, potentially affecting system performance (and possibly ping).

Result

Closing launchers/tabs produced no meaningful change (within typical noise—about <1% change in graphics score).

---

Windows Defender impact test

He tests whether Windows Defender affects the benchmark:

• Disables Windows Defender real-time protection temporarily for the test.
• Notes Defender may reset after reboot.
• Mentions the possibility of more persistent disabling via registry edits, but frames it as optional and dependent on whether performance improves.

Result

Turning off Defender real-time protection produced no improvement in the speaker’s setup.

---

NVIDIA Control Panel tuning (biggest “free” gains)

He recommends two key NVIDIA Control Panel → Manage 3D settings changes:

1. Power management mode → Prefer maximum performance
2. Texture quality → High performance

He claims that for this change, any image quality difference is not discernible.

Result

• A noticeable increase in Time Spy graphics score (from roughly mid-7600s to about 78xx).
• About ~1–1.5 FPS improvement in reported FPS metrics.

---

MSI Afterburner (no “dangerous” overclocking required)

• Highlights MSI Afterburner as a free tool that supports both NVIDIA and AMD, used to adjust GPU sliders/limits.
• Key argument: MSI Afterburner can’t exceed what the GPU BIOS permits in terms of voltage/power limits, making it safer than “scary overclocking” narratives—assuming reasonable settings.

---

Limit/fan/power behavior changes (more performance, less “stable clock dropping”)

He explains GPU Boost behavior:

• Boost frequency depends on power limit, temperature limit, and voltage.
• Frequency increases until one of those limits is hit.

He then adjusts:

• Power limit up to 100%
• Temperature limit to about 83°C
• Fan curve to ramp more aggressively (enabling a user-defined fan profile)

Result

• Further benchmark improvements:
  • Small FPS changes, but enough to move the benchmark score by hundreds of points overall.
• Emphasizes that:
  • Lower temperatures help maintain higher clocks longer
  • Temperature throttling can start relatively early (he cites roughly ~25°C as a “first step-down behavior” point in his explanation, calling it “kind of ridiculous”).

---

Optional “quick dirty” overclocking sliders (still “free,” but not conservative)

At the end, he applies a simple MSI Afterburner overclock:

• Core voltage: 100%
• Core clock: +100 MHz
• Memory clock: +500 MHz

He notes:

• On many GPUs, memory overclocking often has more impact than core clock.
• Results vary by card.
• He also cautions that results may be smaller at 1440p and depend on hardware.

Result

He claims a measurable FPS increase, sometimes larger than earlier steps, but still incremental and hardware-dependent.

---

Overall takeaways

• The biggest “no-cost” improvements come from:
  • NVIDIA Control Panel power management
  • NVIDIA texture quality setting
• Background app/launcher activity and Defender disabling didn’t noticeably help in his specific case.
• Temperature/boost behavior matters: power/temp/fan limit tuning helps maintain clocks longer.
• Optional overclocking can add more performance—often via memory clock—but gains are incremental and vary by GPU.

---

Main speakers / sources

• Speaker: “Jays $0.02” (host of the tutorial)
• Product/game mentioned (not a performance source for the GPU tutorial): World of Warships (sponsorship/code mentioned)