If NVIDIA RTX 5090 was difficult to review, its offer of 80 classes is even more challenging. PC users have become accustomed to this card level that exceeds the outgoing flagship, but let's be clear: RTX 5080 is a slower card than RTX 4090, while the new RTX 5090 is much faster. In many ways, you can think of the new 5080 as a super 4080 Super. There is a small but appreciated increase with the performance against its predecessor, around 13 percent in my tests, and is delivered for the same price. And, of course, you get the last set of DLSS 4 functions, including the generation of multiple frames, which we will see in this review. So, this is the best GPU you can buy at its price, but it is also the only New GPU available at this price.
Today, we are checking the Founders Edition card, which is another Nvidia winner. With RTX 4080 and RTX 4080 Super, industrial design is simply fantastic. The massive and careless design of the latest generation gives way to a new product of 80 classes that once again uses 90 classes cooler. That means you get a relatively small graphic and two slots. The food is supplied by the now standard (for NVIDIA) of 12 VHPWR at the top, while video outputs consist of an HDMI 2.1 port along with three Displayport 2.1S. I had no problem with noise here and similar to RTX 5090, I loved corrugated cardboard packaging.
With a price of $ 999/£ 979, at least for the founder model, I cannot avoid feeling that once again, Nvidia has provided the best model available at the lowest price and, according to the rumors surrounding this launch (together with The third extremely expensive third. One of the key elements that makes this product acceptable is its price: better value than RTX 4090 and RTX 5090 in terms of 'monetary unit per plot' and faster/richest in features than RTX 4080 Super for the same price. Adding a couple of hundreds of pounds or dollars returns us to the dark days of 2022, where RTX 4080 reached $ 1199/£ 1189.
RTX 5090 | RTX 5080 | RTX 5070 TI | RTX 5070 | |
---|---|---|---|---|
Processor | GB202 | GB203 | GB203 | GB205 |
Core | 21,760 | 10,752 | 8,960 | 6,144 |
Impulse clock | 2.41GHz | 2.62 GHz | 2.45 GHz | 2.51 GHz |
Tensor core tops | 3352 | 1801 | 1406 | 988 |
RT nucleus tflops | 318 | 171 | 133 | 94 |
Memory | 32GB GDDR7 | 16GB GDDR7 | 16GB GDDR7 | 12GB GDDR7 |
Memory bus width | 512 bits | Of 256 bits | Of 256 bits | Of 192 bits |
Memory bandwidth | 1792GB/s | 960GB/s | 896GB/s | 672GB/s |
Total graphic power | 575W | 360W | 300W | 250W |
PSU recommendation | 1000W | 850W | 750W | 650W |
Power connector | 600W PCIE 5.0 (4x 8 pins) | 450W PCIE 5.0 (3x 8 pins) | 300W PCIE 5.0 (2x 8 pins) | 300W PCIE 5.0 (2x 8 pins) |
Price | $ 1999/£ 1939 | $ 999/£ 979 | $ 749/£ 729 | $ 549/£ 539 |
Release date | January 30 | January 30 | February | February |
Looking at the specifications, it is easy to see why there is such abyss in the capacities between RTX 5080 and RTX 5090. It is far from the RTX 30 amps line, with 3080 and 3090 they were based on the same processor. In architectural terms, you can almost think of the 5080 processor as half of the 5090s. The 512 -bit interface becomes 256 bits (although 5080 obtains slightly faster GDDR7) and there is a little less than half of the CUDA nuclei . At 360 W, 5080 reaches its maximum point in around 63 percent of the monstrous requirement of 575W of the flagship. Fortunately, the use does not scale linearly with the specifications, so the game's performance gap between them closes.
What the table does not show is how close are the RTX 4080 Super and RTX 5080 points: the core specification points are very similar. The size of the chip is practically identical. RTX 5080 has 10,752 CUDA nuclei against 4080 Super's 10,240. Where there are variations, it is power and bandwidth. RTX 5080 moves to 360W compared to 4080 Super's 320W. The use of GDDR7 for 960 GB/s bandwidth means that there is a 30 percent increase there compared to the outgoing card, so we should see higher performance elevations than the average in the games that benefit from A superior memory bandwidth. Comparisons against 4090 are also interesting: much less CUDA nuclei (10,752 compared to 16,384), but slightly higher watches and only a small reduction in bandwidth, despite the reduction of the memory interface. RTX 4090 will obviously be faster, but the variation in the results is intriguing.
For all our results, we are using a high -end system based on the fastest game CPU, AMD Ryzen 7 9800x3D, to change the load to the graphics card as much as possible. We also have 32 GB of corsair memory DDR5-6000 CL30, a high-end ASUS ROG ROG X870E Hero and a 1000W PSU Corsair hero.
With all that said, let's enter the reference points.
NVIDIA ANALYSIS GEFORCE RTX 5080
- Introduction [This Page]
- RT Benchmarks: Alan Wake 2, Avatar: Frontiers of Pandora, Cyberpunk 2077
- RT Benchmarks: Dying Light 2, F1 24, Hitman: World of Assassination
- RT Benchmarks: Indiana Jones and The Great Circle, Exodus Mayanceded Edition Metro, A plague tale: Requiem
- Reference game: Alan Wake 2, Black Myth: Wukong, Cyberpunk 2077
- Reference game: F1 24, Forza Horizon 5, Senua's Saga: Hellblade 2
- Reference game: Hitman: World of Assassination, A PLANGE TAL: Requiem
- DLSS 4 and route tracking: Cyberpunk 2077, Alan Wake 2
- CONCLUSIONS, VALUE AND RECOMMENDATIONS
To see this content, enable guidance cookies.