GeForce RTX 4080 SUPER is not even super?

 In this video , Artline examines whether the updated RTX 4080 deserves the title of SUPER and how close it comes to the unattainable price and performance of the RTX 4090. They test the GeForce RTX 4080 SUPER against the original RTX 4080 and the flagship GeForce RTX 4090 to see the differences in performance.

Buy RTX 4080: https://artline.ua/search/result?search=RTX+4080+SUPER

Test bench configuration:

• Processor: INTEL Core I9-14900KF
• RAM: G.Skill Trident Z5 Black DDR5-6000 64GB (2x16GB) CL30
• Motherboard: GIGABYTE Z790 AORUS ELITE AX
• SSD: WD Black SN770 1TB
• Cooling system: JONSBO TW7-360 ARGB Black
• Power supply: GIGABYTE UD1000GM 1000W

Video cards for testing:

• GIGABYTE RTX 4080 SUPER WINDFORCE
• MSI RTX 4080 GAMING TRIO
• GIGABYTE RTX 4090 WINDFORCE V2

Text option

“I admit, the GeForce RTX 4080 Super interested me the most from the updated Super line: the RTX 4090 is too expensive, big, hot and power hungry, and the 4080 is still too expensive, but at the same time much weaker than the flagship. Nvidia desperately needed a card in the middle: more powerful 4080 and not for all the money in the world.

Reducing the price from $1,200 to Nvidia's thousand was easier than letting mere mortals touch their big AD102 chip and create a 4080Ti.

Instead, the greens took the simple route: they left the same AD103 chip as in the original 4080, only without damage. After the 4080 is cancelled, damaged crystals will go into service in the 4070 Ti Super. And if it seemed to you that the full-fledged AD103 in the new product does not give so many bonuses, then it didn’t seem to you: 5 and a penny percent of the addition of CUDA, tensor, RT cores and texture units is unlikely to excite enthusiasts. Nvidia also understands this, so they also accelerated the GDDR6X video memory by as much as 2.7%. GPU: AD103-400-A1 (TSMC 5 nm) up to 2550 MHz

• 45900 million transistors
• Shading Units: 10240 (+5.2%*)
• TMU: 320 (+5.2%*)
• Tensor/RT cores: 320/80 (+5.2%*)
• ROP: 112 (0%*)
• L2 Cache: 64 MB (0%*)
• VRAM: 16 GB GDDR6X, 256 bit, 23 Gb/s (+2.6%*)
• TDP: 320 W (0%*) *compared to RTX 4080

Nvidia is embarrassed to compare this huge, in quotes, boost with the original 4080, but compares it with the previous generation in 4K.

I won’t be shy and will compare the super-new product with the RTX 4080 and the flagship RTX 4090 in modern games with 4K resolution on the configuration specified in the description of the video.

Assassin's Creed Mirage, 4K, highest graphics settings without upscale and dynamic resolution. A five percent increase in the number of CUDA cores brought in practice only 2 and a penny percent and an increase in the average and minimum framerate. Well, yes, the beginning is not impressive. The RTX 4090 is more than a quarter more powerful than the new product, and this is in a game that does not use ray tracing and is not the most demanding. RoboCop: Rogue City, 4K, Epic graphics preset, with high quality Lumen reflections enabled. The game is very demanding on maximum settings, especially in gunplay. The 4080 Super is not impressive: one frame and one and a half percent difference compared to the original 4080 is not enough for the Super title. Call of Duty: Modern Warfare III, 4K, Extreme graphics settings preset. Owners of high-Hz 4K monitors can play with 4080 and 4090 even without DLSS, and the updated built-in benchmark has much less stuttering. However, problems with smoothness still remained and affected the original 4080. Otherwise, no miracles happened: plus 2 percent FPS in the super update and the 4090 is unattainable. Starfield, 4K, Ultra graphics settings, no upscale. 4080 and 4080 Super are not enough to confidently keep the FPS bar at sixty, but still DLSS clearly suggests itself here. The RTX 4090, as in previous games, feels like a duck to water, outperforming its younger sisters by almost 30 percent and having headroom for 60Hz monitors and TVs.

Star Wars Jedi: Survivor, 4K, epic graphics settings with RT. The gap between the 4090 and 4080 Super is still huge. The difference of 37 percent and a penny looks sad against the backdrop of prices for the 4090 and old rumors that we will see the 4080 Ti in the first quarter of this year. But in the end we got a 3 percent increase compared to the 4080, hello DLSS and compromises for a thousand dollars.

Hogwarts Legacy, 4K, ultra settings with ray tracing enabled. Hogwarts again did not want to work on both 4080 without DLSS, you see, with 16 gigabytes of video memory it is uncomfortable. And with DLSS the emphasis begins to be unclear. As a result, the RTX 4090 broke away from the other cards by only 12-13 frames - not the best reason to overpay more than a thousand dollars in a game that has not been optimized in a year.

Well, what would we do without Cyberpunk 2077. 4K, RT Overdrive preset with ray tracing, DLSS Quality. There is, after all, a game that brings absolutely all video cards to their knees, including the RTX 4090. There is no way to generate frames here. Well, the new product turned out to be faster than its predecessor by as much as one and a half percent - that’s less than a frame per second. The 4090 is almost 40 percent faster, but the frame rate is still not impressive. As for the operation of the cooling system, power consumption and temperatures, I won’t say much, since there are no major changes compared to the RTX 4080.

The noise can be described as average, but you already understand that no one except Noctua has been able to make 320-watt cards silent. By the way, about the 4080 from Noctua there is a video about a silent PC on the channel. In general, the 4080 Super is not Ti at all, and we are unlikely to get Ti and it’s clear why: the large AD102 chip continues to print millions in green, which China and moneybags are willing to give away. Even the 4090D with the same price tag is rejected with a bang.

But there is still something positive now: even at the start of sales, the super-version 4080 will be cheaper than the original 4080."