The largest companies on the market have a series of products that allow them to completely dominate certain sectors, in the case of video games Nvidia He has chart This offers very high performance being the main choice of players. But they also have models designed for professionals who need more power for certain applications, and although they have better specifications and are more expensive, they generally do not have good performance in other aspects such as games.
One of the main reasons why large companies in the IT sector have a lot of products have to do with user needs, the problem that there is no equipment combined with generalized software to offer the best performance in all aspects. Graphics cards, for example, need periodic drivers to be up to date with all programs, the difference is that those used for games very often change certain aspects while those designed for professionals try to maintain stability to be compatible with all possible programs.
Can an extremely expensive graphic designed for professionals be used to play?
Surely everyone knows what a GeForce graphics RTXBut there are not many users who know that there is a much more expensive range which has much higher specifications but which has a very different target audience. In this case, we are talking about models Quadro RTX (Now called as RTX PRO), a type of Gpu It takes a long time on the market and which are generally in computers such as the workstation focused on rendering, AI and other aspects, but you will never see on a computer designed for games.
It may first seem strange when we see the characteristics they have, if for example we compare the last quadro model which was launched with the best GeForce graphic of that time, we will see that he has superior specifications, something similar to what we compare the last model RTX PRO With the best GeForce RTX real:
| Specifications | Quadro RTX 8000 | GeForce RTX 3090 | RTX 6000 Pro | GeForce RTX 5090 |
|---|---|---|---|---|
| Núcleos cuda | 4608 | 10496 | 24.064 | 21,760 |
| Rops | 96 | 112 | 176 | 176 |
| Tensor nuclei | 576 | 328 | 752 | 680 |
| RT | 72 | 82 | 188 | 170 |
| Sms | 72 | 82 | 188 | 170 |
| Clock speeds | 1395 MHz 1770 MHz |
1395 MHz 1695 MHz |
4000 | 3352 |
| Tflops fp32 | 16 | 36 | 125 | 105 |
| Truum | 48 GB GDDR6 | 24 GB GDDR6 | 96 GB GDDR7 | 32 GB GDRR7 |
| Memory interface | 384 bits | 384 bits | 512 bits | 512 bits |
| Bandwidth | 672 GB / s | 936.2 GB / s | 1792 GB / s. | 1792 GB / s. |
| TDP | 260 W | 350 W | 600 W | 600 W |
Its basic performance is still lower than a RTX 3090 would be a model that costs more than 2000 $, which is mainly due to the fact that it does not aim to offer good game performance because it is designed for workloads related to rendering or AI. The main difference between an RTX and a quadro is, in addition to the price and specifications, in the pilots they use. While RTX uses ready to play controllers (designed for games), the Quadro uses studio drivers (created for creative applications).
Even if, it is possible to use them for certain games, although it is obviously in those who use the CPU more than the GPU, because obviously, the fact of not having the necessary controllers are not optimized for this, the actual manner which should be able to use it would be to optimize the processor as much as possible so that it is responsible for the execution of the heaviest.
