NVIDIA’s Flagship GeForce RTX 4090 graphics card A brief gaming demo featuring Cyberpunk 2077 and DLSS 3 showed off some huge numbers.
NVIDIA GeForce RTX 4090 Cyberpunk 2077 Demonstrated: 2x Boost with DLSS 3 in Stock, 2.85 GHz Clocks, Temperatures Around 50C and Low Wattage
The demo shown was using an NVIDIA GeForce RTX 4090 Founders Edition graphics card that we fully detailed here, The card was paired with an Intel Core i9-12900K system and running Cyberpunk 2077 at 1440p using Ultra Ray Tracing and Psyco settings (or maximum settings). At its native resolution, without DLSS and Reflex, the game ran at an average of 59 FPS (49 FPS down 1%), given a fixed frame and 72-75ms latency.
At these settings, the GPU was running over 2.8 GHz, averaging around 2810-2850 MHz (min/max), and with 100% usage, the temperature remained stable between 50-55C. That’s a reference boost clock difference of 330MHz vs 2520MHz (+13% increase) and the impressive part is that there was no overclocking involved! All this was happening in stock. This is just one game and we may see even higher clock speeds in other games. the card has since informed of To run over 3 GHz with overclocking.
But now we have to talk about performance dllss 3 enabled, The game was using a pre-release version of DLSS 3, so performance and settings will be different in the final version. As soon as the DLSS 3 setting is turned on, the DLSS frame generation setting is also enabled. It was using the “Quality” preset and we once again saw full GPU usage with clocks over 2.8GHz, but this time the temperature was closer to 50C than 55C (keep that in mind). The NVIDIA GeForce RTX 4090 boosted performance to an average of 170 FPS (119 FPS less than 1%) with DLSS 3 enabled and an average latency of 53.00ms. This is a 2x improvement in FPS and a 30% improvement in latency reduction versus DLSS disabled.
NVIDIA GeForce RTX 4090 (DLSS 3 disabled) in Cyberpunk 2077:
NVIDIA GeForce RTX 4090 (DLSS 3 Enabled) in Cyberpunk 2077:
But not only that, using the latest PCAT tool, which also comes with support for PCIe 5.0 16-pin (12VPHWR) power plugs, NVIDIA also provided wattage figures with both DLSS disabled and enabled. With DLSS 3 disabled, the NVIDIA GeForce RTX 4090 graphics card consumes an average of 461 watts with a performance of 0.135 points per watt (Frames/Joule). As soon as DLSS 3 was enabled, the GPU saw wattage as low as 348W, or a 25% reduction. It also increased the perf per watt by 0.513, an increase of 3.8x.
NVIDIA GeForce RTX 4090 Cyberpunk 2077 Demo
|Cyberpunk 2077 1440P (Max Settings + Ultra RT + Psycho)||NVIDIA GeForce RTX 4090||NVIDIA GeForce RTX 4090||NVIDIA GeForce RTX 3090 Ti||NVIDIA GeForce RTX 3090 Ti|
|DLSS 3||Disabled (Basic)||able (quality)||Disabled (Basic)||Enabled (Quality DLSS 2)|
|FPS (1% less)||49.3||119.6||,||,|
|GPU Clock||2800-2850MHz (Stock)||2800-2850MHz (Stock)||2000-2050 MHz||2000-2050 MHz|
|PCAT Perf/Watt (FPS/Joule)||0.135||0.513||0.079||0.152|
Updates: We’ve also added performance results from our GeForce RTX 3090 Ti with DLSS enabled and disabled. We’re running a custom MSI SUPRIM X design that has a higher factory OC, so performance should be slightly better than the FE version. The card in stock scored an average of 37.2 FPS at native 1440p under the same settings and the same scene. The card consumed 452W of power and is running around 75C. With DLSS 2 “Quality” enabled, we saw a boost of 61 FPS (+64% increase) and power draw also dropped to 409W (-100% decrease). Following is the result.
Cyberpunk 2077 Ultra Quality + Psycho RT (original 1440p):
- MSI RTX 3090 Ti Superim X (Stock Native 1440p) – 37 FPS / 455W Power / ~ 75C
- Nvidia RTX 4090 FE (Stock Native 1440p) – 60 FPS / 461W Power / ~ 55C
- RTX 4090 vs RTX 3090 Ti = +62% faster
Cyberpunk 2077 Ultra Quality + Psycho RT (DLSS 1440p):
- MSI RTX 3090 Ti Superim X (DLSS 2 1440P) – 61 FPS / 409W Power / 74C
- Nvidia RTX 4090 FE (DLSS 3 1440p) – 170 FPS / 348W Power / ~ 50C
- RTX 4090 vs RTX 3090 Ti = +178% faster
The power number is a seriously big deal and one of the reasons this might happen is that the load of the FP32 cores goes to the tensor cores that run the DLSS algorithms. These cores are specialized in these negotiations and instead of forcing the entire GPU, resulting in a higher power draw, tensor cores can process data much faster and more efficiently, thereby reducing power consumption. The DLSS 3 can be a game changer in power efficiency and performance efficiency and we can’t wait to really test it out when we get our review samples.
Products mentioned in this post