We all desire NVIDIA’s RTX 4090, but it’s difficult to accept either its high power requirements or its $1,599 starting price (assuming you can even find it at that price). The RTX 4080 with 16GB of VRAM, which costs $1,199, is the only other new NVIDIA choice left for anxious PC gamers this year. In the context of high-end PC gaming, $400 isn’t exactly a large savings (particularly not compared to the $899 12GB RTX 4080 that NVIDIA “unlaunched”), but it might be enough to entice certain players.
It works with NVIDIA’s potent new DLSS 3 upscaling technology, and it is quicker than the RTX 3080 Ti that was introduced at the same price earlier this year (which is limited to 4000-series GPUs). The RTX 4080 is a potent GPU that will please anyone who wants to game in 4K with ray tracing if they can live without the ego boost of owning a 4090. You’re definitely better off holding out for AMD’s anticipated 4070 and 4060 cards as well as its planned RDNA 3 GPUs if you’re stuck with lower-resolution monitors.
Surprisingly, the RTX 4080 Founders Edition we evaluated and the 4090 have the same layout. Both of them have enormous vapour chambers, three PCI-e slots, and the original pass-through fan architecture from NVIDIA’s earlier GPUs. To be honest, I had anticipated something a little smaller. The 4080, in contrast to the 4090, only requires three 8-pin PSU cables to operate. (Both cards can also be supplied by a single PCIe 5.0 PSU line, although at the present, those power supplies are quite uncommon.)
The power connections of the 4080 also allude to one of its main benefits: It has a 320-watt thermal design profile (TDP) and needs a 750-watt power supply, compared to the 4090s much higher 450-watt TDP. Upgrade to the 4090 may require purchasing a new unit and reconnecting electricity throughout the complete system unless you already have an 850W power supply. Although these cards won’t always operate at their highest power loads, you still need to be prepared for the infrequent occasions when they do.
Even while it shares a similar exterior to the 4090, the RTX 4080 is a very different beast within. 9,728 CUDA cores, 16GB of GDDR6X VRAM, and a base clock speed of 2.21GHz power it (with boost speeds of 2.51GHz). The 4090, on the other hand, features a staggering 24GB of VRAM and 16,384 CUDA cores with somewhat faster clock rates. With NVIDIA’s new Ada Lovelace architecture, noticeably better performance, and 4GB more VRAM than the 3080 Ti, the 4080 triumphs over it. Although the 3080 Ti officially has about 500 more CUDA cores than the new NVIDIA platform, these cores are slower and less effective by nature.
What do these figures actually signify, then? In comparison to the 4090, the RTX 4080 scored roughly 3,500 points less in 3DMark’s TimeSpy Extreme benchmark. However, the 4080 would be the most proficient GPU we’ve ever examined if that more potent card didn’t exist. It achieved a comfortable 130 frames per second while playing Halo Infinite in 4K with all of its graphic settings set to their highest levels, with its TimeSpy Extreme score being around 50% higher than the 3080 Tis. With the aid of DLSS 3, Cyberpunk 2077 was able to achieve 74 frames per second in 4K. This nearly made me cry.
A word about DLSS 3: It is NVIDIA’s most recent AI solution that has the ability to upscale lower-quality pictures to higher resolutions. However, DLSS 3 may also add interpolated frames to smooth out 4K gaming in addition to intelligently sharpening edges and improving textures. While testing Cyberpunk and A Plague Tale: Requiem with the technology active, I didn’t notice any strange framerate hiccups, despite the fact that I can occasionally spot problems with extremely low-quality DLSS upscaling.
The RTX 4080’s only significant drawback is that I can’t help but contrast it with the 4090. On the 4090, the same Cyberpunk ray tracing benchmark ran at an incredible 135 frames per second. Additionally, it achieved a 40 fps increase in average framerate in the 3DMark Port Royal ray tracing test. However, these are the kinds of improvements that only the most devoted players will see, making them the ideal market for the 4090. Even with ray tracing in taxing games like Control, the RTX 4080 never made me feel limited during genuine 4K gameplay.
And overclocking is always an option if you want extra performance. Although I didn’t get a chance to test it personally, the 4080’s thermal performance leads me to believe there is lots of space for improvement. During my testing, it never got over 61 degrees Celsius, which is roughly 10 degrees lower than the 4090. That is evidence of NVIDIA’s superior cooling system (and perhaps partially due to my office being slightly cooler this month).
Is it worthwhile to accept the 4080 if there’s a chance you would be able to purchase the 4090 for $1,599? Currently, the majority of online shops charge significantly more than $2,000 for 4090 cards. With that gap, the $1,199 card appears like a great deal, which sounds absurd to suggest. But nobody knows how long the RTX 4080 will remain available at its launch pricing. It won’t be long before it gradually approaches the higher price of the 4090.
Additionally, if you think it’s ridiculous to pay more than $1,000 for a video card — and let’s be clear, you should – wait to see what NVIDIA’s upcoming cards will look like. The question is when, but we do anticipate RTX 4070, 4060, and 4050 GPUs in the future. What the hell is NVIDIA going to accomplish with its $899 4080 GPU? Is it going to be the 4070? The top RDNA 3 GPUs from AMD will be available for less than $1,000, and Intel’s new Arc GPUs are unexpectedly strong at the base level.
Overall, the RTX 4080 is all I could have asked for in an RTX 3080 Ti replacement. It is speedier and has a tonne of new functions, making it a clear improvement over the prior cards. I wouldn’t blame you for being tempted by the 4080 if you came into $1,199, but I’m not saying you should upgrade your 3080 any time soon.
