A modified NVIDIA GeForce RTX 2080 Ti equipped with 22GB of GDDR6 memory has delivered up to 73% higher gaming performance than the original 11GB model in recent benchmarks.
The results demonstrate how additional video memory can improve performance in demanding games, particularly at 4K resolution.
The RTX 2080 Ti originally launched in 2018 with 11GB of GDDR6 memory. The modified version doubles that capacity by replacing its original 1GB memory chips with 2GB Samsung modules.
Testing revealed substantial improvements in several modern games, although others showed almost no difference.
RTX 2080 Ti 22GB Shows Major Improvements at 4K
The modified graphics card was tested at 1440p and 4K using the same GPU architecture as the standard RTX 2080 Ti.
At 1440p with DLSS Quality enabled, performance improvements were generally limited.
Microsoft Flight Simulator 2024 recorded a 29% increase in average frame rates, while Spider Man 2 improved by approximately 22%.
However, Cyberpunk 2077 and Gears of War: E Day showed no measurable improvement at this resolution.
The differences became considerably larger at 4K, where higher resolution textures and other graphical resources increased memory requirements.
Microsoft Flight Simulator 2024 improved from approximately 30 FPS to 52 FPS, representing a 73% increase.
Dragon Age: The Veilguard increased from 16 FPS to 27 FPS, while Spider Man 2 improved from 39 FPS to 64 FPS.
RTX 2080 Ti 11GB vs 22GB Gaming Benchmarks
| Game | 11GB at 4K | 22GB at 4K | Improvement |
|---|---|---|---|
| Flight Simulator 2024 | 30 FPS | 52 FPS | 73% |
| Dragon Age: The Veilguard | 16 FPS | 27 FPS | 68% |
| Spider Man 2 | 39 FPS | 64 FPS | 64% |
| Gears of War: E Day | 14 FPS | 17 FPS | 21% |
| Cyberpunk 2077 | 42 FPS | 48 FPS | 14% |
| Assassin's Creed Black Flag Resynced | 60 FPS | 62 FPS | 3% |
| Pragmata | 72 FPS | 73 FPS | 1% |
The 4K tests used DLSS Performance and FSR Frame Generation, where supported.
Additional VRAM Also Benefits AI Workloads
The modified RTX 2080 Ti also demonstrated improvements in certain AI workloads, where larger models and working datasets can require more memory.
Having 22GB available reduces the need to transfer data between system memory and GPU memory when workloads exceed the original card's capacity.
However, additional VRAM does not increase the GPU's underlying processing capability or memory bandwidth.
This explains why games that already fit comfortably within 11GB showed minimal improvements.
The experiment highlights the growing importance of graphics memory capacity for modern gaming and AI applications.
Nevertheless, the results should not be interpreted as evidence that modifying older graphics cards is practical for everyone. Memory upgrades require specialized equipment and can introduce compatibility or reliability problems.
The benchmarks primarily demonstrate how significantly insufficient VRAM can limit performance, even when the GPU itself remains capable of running demanding workloads.
Discussion (0)
Be the first to comment.