We've tried to strike a balance between games with effects and games with ray tracing without it, although most AAA games these days include it, and modern GPUs should be able to handle it just fine (best of luck to AMD with your next RDNA 4 cards).
For the 5090, we have run all tests in 4K, if you don't mind running games in 4K, even if you want super high frame rates at 1440p or for some type of ultrawide monitor, the 5090 is probably overkill. When we run high-scale tests, we use the newest version of DLSS available for NVIDIA cards, the latest version of FSR available for AMD cards, and the latest version of Xess available for Intel cards (not relevant here, just stating for the record ), and we use the “quality” setting (in 4K, which is equivalent to a true rendering version of 1440p).
Rendering performance: much faster, much more power hungry
Before we talk about frame generation or “fake frames,” let's compare apples to apples and just examine the 5090's rendering performance.
The card primarily benefits from four things compared to the 4090: the updated Blackwell GPU architecture, a nearly 33 percent increase in the number of CUDA cores, an upgrade from GDDR6X to GDDR7, and a move up a memory bus. from 384-bit to a 512-Bit Bus. It also jumps from 24GB of RAM to 32GB, but games still generally don't rush to a 24GB limit, so So increasing the capacity itself shouldn't really change performance if all you're focused on is gaming.
And for people who prioritize performance over everything else, the 5090 is a big deal: It's the first consumer graphics card from any company that's faster than a 4090, since Nvidia never stocked the 4090 last year when it made its generation generation Super supercenration. Update of the 4080, 4070 Ti and 4070.
Comparing games rendered natively at 4K, the 5090 is between 17 percent and 40 percent faster than the 4090, with most games we tested landing somewhere in the low to high 30 percent range. . That's an undeniably big hit, one that's roughly in line with the increase in CUDA core count. Tests run with DLSS enabled (offset only and with frame generation in 2x mode) improve by about the same amount.