To be taken with caution: First Timespy value of the Geforce RTX 4090 [Gerücht]

According to the publication of “kopite7kimi”, which is known for its hardware leaks, Nvidia’s upcoming flagship graphics card allegedly achieves more than 19,000 points in the well-known 3DMark Time Spy Extreme benchmark. A comparison between the three top models of the current “Ampere” and the upcoming “Ada” generation suggests that the RTX 4090 is 66 percent faster than the RTX 3090 Ti and 82 percent faster than the RTX 3090 in this benchmark alone , reported Videocardz.com.

The RTX 4090 is rumored to have 16,384 CUDA cores, which is 52 percent more than the RTX 3090 Ti. Accordingly, increased clock rates and increased power consumption, as is already expected, should also play a role in the result, provided the leak correct is. In other tweets, “kopite7kimi” also says that the RTX 4000 series can easily run at over 3 GHz, while the RTX 4090 is clocked lower. According to his assessment, an overclocked RTX 4090 can also achieve over 20,000 points in the said benchmark.

To be taken with caution: First Timespy value of the Geforce RTX 4090 [Gerücht]

Source: videocardz.com




Meanwhile, the cache size is said to play an important role in the coming generation: “Because the size of the cache has a direct effect on the bandwidth, which in turn has a significant impact on performance.” Since this is not yet official or comprehensible data, the whole thing should be treated with caution, especially since the performance plus can look different in other benchmarks and games. While there have been repeated rumors of a postponement to the beginning of next year due to high stock levels, the RTX 4090 as the alleged first offshoot of the new generation is not expected until October at the earliest.