TecLab has re-published the review of the new RTX before its official premiere. This time the network included tests of the most powerful Ampere chip – the RTX 3090. According to their review, the RTX 3090 is on average only 10% faster than the RTX 3080 (being twice as expensive). This is certainly due to a bottleneck, but the results are very interesting and at the same time highly controversial. TecLabs aspires to be one of the best leak sources and testers wearing husky masks don’t care about the Nvidia embargo, posting richly detailed material sooner than anyone else. This time it is the RTX 3090 review which surprisingly presents only 10% more performance than the RTX 3080.
The RTX 3090 is only 9% faster than the RTX 3080. The problem is the software is not able to use the power of the chip.
| NVIDIA GeForce RTX 3090 vs RTX 3080 | |||
|---|---|---|---|
| Result / 4K FPS | RTX 3090 | RTX 3080 | 3090/3080 |
| 3DMark Time Spy Extreme | 9948 | 9000 | + 10,5% |
| 3DMark Port Royal | 12827 | 11981 | + 7,1% |
| Metro Exodus RTX / DLSS OFF | 54.4 | 48.8 | + 10,2% |
| Metro Exodus RTX / DLSS enabled | 74.5 | 67.6 | + 10,2% |
| Rainbow Six Siege | 275 | 260 | + 5,8% |
| Horizon Zero Dawn | 84 | 76 | + 10,5% |
| Forza Horizon | 156 | 149 | + 4,7% |
| Far cry | 107 | 99 | + 8,1% |
| Assassins Creed Oddysey | 71 | 65 | + 9,2% |
| Shadow of the Tomb Raider RTX / DLSS wyłączony | 91 | 83 | + 9,6% |
| Shadow of the Tomb Raider RTX / DLSS włączony | 111 | 102 | + 8,8% |
| Borderlands 3 | 67.6 | 61.3 | + 10,3% |
| Death Stranding DLSS / RTX włączony | 175 | 164 | + 6,7% |
| Death Stranding DLSS / RTX WYŁĄCZONE | 116 | 104 | + 11,5% |
| Control DLSS / RTX ON | 71 | 65 | + 9,2% |
| Control DLSS / RTX OFF | 62 | 57 | + 8,8% |
–
The tests were carried out using an Intel Core i9-10900 processor clocked at 5 GHz with 32 GB of 4133 MHz RAM, which is a very solid set for such a task. In addition, the Galax HOF PRO M.2 1 TB SSD. While reviewers aren’t talking directly about the RTX 3080 and RTX 3090, they’re talking about the flagship models of 5000 YUAN and 10000 YUAN, so it’s clear what they’re referring to. The following data was collected by Videocardz. As you can see, the average performance gain in a panel of 16 synthetic tests and games is about 9% less compared to the RTX 3080. These results suggest that something is wrong here. However, given that the RTX 3080 already doubles the number of CUDA cores compared to the RTX 2000 series, any limitations in the software will appear in this case, not to mention the RTX 3090.
Why is the RTX 3090 only 9% faster than the RTX 3080? Probably the cores spike in the 3000 series is just too big for the current software to take advantage of it. Although the drivers would (likely) be updated to handle the enormous bandwidth, the game code and engines need to be scaled up to use the available processing power. Unfortunately, you can see problems with scaling games here and the software itself is clearly the bottleneck here. However, when there are positions that can harness this enormous power, the RTX 3090 will show what it can do.
– .
