Technology News – Nvidia's top-of-the-line Blackwell architecture-based card is said to be capable of clock rates that we haven't seen many of in video cards to date.
At the Chiphell forum, Panzerlied, who previously provided detailed information about Nvidia's GeForce RTX 5000 graphics cards, mentioned something again, but this time about the top model. Officially, the “green” did not say much about the new generation, but company president Jensen Huang said during a Q&A session at Computex that they will be special products and he can't wait to say more about them. But what makes the RTX 5090 so muscular?
According to rumors, the clock speed of the video card will be around 2.9 GHz by default, so it will not be far from 3 GHz. Here it is worth comparing with the current flagship, the RTX 4090, which operates at 2235 MHz by default, so we can report a 30% improvement in this regard alone. However, we do not yet know the magnitude of the boost clock signal, but given the current generation, it will definitely be more than 3 GHz (the RTX 4090 has 2520 MHz on the reference card, but according to individual manufacturers, it is 2.8 or even 2.9, and maybe even 3 GHz with the appropriate consumption and cooling system).
The GeForce RTX 5090 will likely use the Blackwell GB202 GPU with a slightly lower-end chip, as it originally had a 512-bit memory bus and 192 SMs (streaming multiprocessors), but that will be scaled down to 448-bit and 28GB of GDDR7 VRAM will be present on the card. Of course, this could still change, and possibly downwards, as Nvidia hasn’t really released a true top-end chip in the past few generations (they usually end in a 0, the GB202 doesn’t…), so they hold us back from what the technology can really do.
The GeForce RTX 5090 and 5080 are said to be released at the end of the year, but it's also possible that Nvidia will delay the launch to early 2025. That won't be so lucky: both Intel and AMD will be launching their new processors soon!
