When NVIDIA introduced its new GeForce RTX 40 series graphics cards and Ada Lovelace GPU architecture in September, it was widely believed that these next-generation GeForce cards would provide PC gamers with previously unattainable levels of performance. And in fact, it is what we have observed.
Gaming Performance
With our synthetic benchmarks as a starting point, the 4090 didn't simply show up; it crashed through the wall like the Kool-Aid Man high on his juice. The 4090 achieved an astonishing score of 21,872 in 3DMark Fire Strike Ultra.
To put that into perspective, it is around 50% higher than the RX 6950 XT's score of 14,512, which was the highest we have ever achieved in-house. It is also roughly equal to the previous overclocking world record for Fire Strike Ultra for not just one but two GPUs simultaneously.
The 4090 outperformed every card we've ever tested, with improvements over the 3090 Ti of 26% at 1080p, 39% at 1440p, and 31% at 4K.
The 4090 nearly doubles, and in some cases even images, the scores of the 3090 Ti in the ray tracing synthetics, continuing the trend. These figures are pretty remarkable.
As we go on to our gaming benchmarks, the 4090 maintains its lead. Borderlands 3, Gears Tactics, Metro Exodus, and Total War: Three Kingdoms make up our typical four-game test suite. If available, Ray tracing and DLSS are enabled, and the highest graphics setting is used for all tests.
Design & Details
It has three compartments and weighs a little over five and a half pounds. However, it is large and gorgeous, using the same silver and black color palette as Nvidia's 30-series cards released two years ago.
Even compared to the still outstanding RTX 3090 Ti, Nvidia's developers created an impressive feature sheet. It has 16,384 CUDA cores, an increase from the RTX 3090 Ti's 10,752 cores. With a base frequency of 2.23GHz that can be upgraded to 2.52GHz and 24GB of GDDR6 VRAM, it has an increase in computational power of 83 shader teraflops from the 3090 Ti, 1,321 tensor teraflops from the 3090 Ti and 191 ray tracing teraflops from the 3090 Ti.
The outcomes Nvidia has produced are substantial, and the boost we've witnessed in most of my testing is unquestionably more than just a slight improvement. This is a significant generational advance.
This card is made to rule 4K gaming, but creators and experts should find its 24GB of GDDR6X video memory very appealing. Even the most demanding games today would find it overkill, but if you're making 3D models, that capability will speed up rendering. With 24GB, OctaneRenderer will rarely need to use the system's RAM and will instead use the card's quicker frame buffer, significantly speeding up rendering. Similar to how you can work in the viewport while your render completes in Blender, you are free to do so.
Moreover, these programs work much better when used. Owing to the GPU's AI processors, denoising renderings occur significantly more quickly.
Price of the RTX 4090
The RTX 4090 is at the absolute top of the GPU market and costs $1,599; nevertheless, some PC cases may find it challenging to fit due to its large footprint. You'll need plenty of extra PCIe power connectors to power the RTX 4090 because it consumes 450 watts.
Which graphics card is the most powerful?
With 640 Tensor Cores and 12 GB of HBM2 memory, the NVIDIA TITAN V is capable of 110 TeraFLOPS speed. For better results, it also has NVIDIA CUDA tuned for Volta.
Conclusion
The RTX 4090 may be large and pricey, but holy smokes if it doesn't outperform the competition. That's a little unfair because there are presently no other cards from this new generation accessible, so we can only compare them to cards from the last several years. But, until the rest of the pack catches up, with its fantastic hardware specs and DLSS 3 AI trickery, even the $1,599 price tag doesn't seem outrageous for the unprecedented frame rates that this card can provide.
