Products You May Like
Nvidia’s RTX 4080 16GB will officially launch next week on Nov. 16. As the slightly more affordable 40-series GPU, there’s strong interest in its gaming chops. Now some preliminary benchmarks have been released showing it could indeed be a very powerful GPU. You might not be able to buy one, but it’s still interesting to see where it stands. It’ll be even more interesting to see how it fares against AMD’s RDNA3 GPUs. Sadly those aren’t launching until Dec. 13, so we’ll need to be patient.
The RTX 4080 16GB benchmarks were posted by Videocardz, which didn’t reveal their sources. They only note they have acquired them. They also posted overclocked results, which look promising. But as always, your mileage may vary. The site notes it tested a “semi-custom” model with an Nvidia reference design PCB. They didn’t post any photos of the board, though, so it’s unclear which model they tested. The card was tested at stock settings, then with 10W added to its power limit, and then overclocked.
Overall, overclocking offered a nine percent boost in 3DMark TimeSpy, and a seven percent boost in Port Royal. That’s nothing to scoff at, as typically overclocking an Nvidia GPU beyond what its own mechanisms are capable of offers minimal gains. That might be different with the 40-series and their massive coolers, though. The RTX 4080 16GB Founders Edition is the same size as the 4090. Add-in boards (AIBs) from partners will likely be equally massive.
Its graphics score of 15,062 is approximately 22 percent slower than the RTX 4090. However, it’s also a surprising 69 percent faster than the original RTX 3080 10GB. That model sold for around $700 if you don’t consider pandemic pricing, whereas the 4080 is going for $1,199. Therefore, a healthy uplift is expected. That’s about the same leap Nvidia made going from the RTX 2080 to the 3080.
In Port Royal, which is a ray tracing benchmark, the uplift was a bit less compared with the RTX 3080, but still decent at 52 percent. This test also shows what a beast the RTX 4090 is, as it’s 40 percent faster than the stock RTX 4080. Nvidia clearly left a yawning gap between these two cards in order to fit an RTX 4080 Ti in there.
Chiphell published more benchmarks posted for popular games, with an interesting data point on the RTX 4080 versus the Radeon RX 7900 XTX. When AMD was talking up its GPUs’ ray tracing prowess, it provided some benchmarks. One of them showed Cyberpunk 2077 obtaining 62fps at 4K with RT set to Ultra. The Chiphell numbers show the 4080 hitting 82fps at the same settings. The Cyberpunk numbers from Chiphell were achieved with a Ryzen 7 7600 X CPU also. It includes 4K benchmarks for Borderlands 3, Forza Horizon 5, Shadow of the Tomb Raider, and CoD MW2 as well. Unfortunately, we don’t have numbers for AMD’s new GPUs to compare.
The RTX 4080 16GB is shaping up to be a great GPU for 4K gaming. Recall that Nvidia released its own RTX 4080 16GB (and 12GB, RIP) benchmarks previously. The purpose of that PR campaign was to show the benefits of DLSS 3, which is exclusive to the 40-series for now. AMD also noted that its own FSR technology is responsible for huge performance boosts for its upcoming GPUs. It’s also targeting 4K gamers (and even 8K) with its RDNA3 GPUs. AMD has said its flagship XTX GPU is aimed directly at the RTX 4080 16GB, even though it costs $200 less.
What remains to be seen about the RTX 4080 16GB is whether it’s truly only a “4K-only” GPU like the RTX 4090. That flagship model can run into CPU bottlenecks even at 1440p with a high-end CPU like the Core i9-12900K. That makes it pretty useless for anyone without a 4K monitor. This might not be the case with the less-powerful 4080, as its AD103 die is almost half the size of that in the RTX 4090. It has fewer CUDA cores and a narrower 256-bit memory bus as well.