The first comparison between the regular Nvidia GeForce RTX 5090 graphics card and the stripped-down model GeForce RTX 5090D has appeared.
Surely, as fans and readers of hardware news, you have known for some time that the Chinese market receives “cut down” models of Nvidia’s powerful graphics cards, the reason being strategic security on a global scale. Last year, almost immediately after the announcement of the series GeForce RTX The 40XX has found comparisons with the GeForce RTX 4090D cards headed to the Chinese market. The “D” version had a reduced number of cores and about 5% performance.
The same will be the case with the recently introduced GeForce RTX 50XX series. Now they come first comparison between RTX 5090 and RTX 5090D versions. This time too, cards with slightly reduced performance will reach the Chinese territory, but players should not feel the change.
Specifications |
RTX 5090 |
RTX 5090D |
| GPU Matrice | GB202 | GB202 |
| CUDA cores | 24 576 | 21 760 |
| Memory | 32 GB GDDR7 | 32 GB GDDR7 |
| Memory speed | 28 Gbps | 28 Gbps |
| Memory bandwidth | 1792 GB/s | 1792 GB/s |
| TDP (Thermal Design Power) | 500 W | 450 W |
| Multiple GPU support | Again | Yes |
| Focus | Gaming and professional workload | Gaming, limited AI applications |
| Availability | Global | Only in selected markets |
However, those who planned to use multiple GPUs at the same time, for example for use in AI clusters or data centers, will be dissatisfied.
Source: pctuning.cz
