On the 12th of this month, Nvidia unveiled the 4000 series graphics card in the market. Of these, the RTX 4090 card is being discussed the most. Today its architecture, specifications, and new technologies will be discussed in detail.
There have been several major changes in the card’s memory configuration. A level one of 18 MB per SMA with a total of 128 kilobytes is provided, along with a huge level two cache of 96 megabytes, which is 6 times more than the previous generation. Tied up with Micron, Nvidia has given Ada Lovelace 24 Gbps speed GDDR6X memory.
Ray tracing performance and upscaling performance will increase by 2 to 4 times due to the new RT core and Tensor core. The main reason for this performance uplift is the change in working principles and internal architecture of these new cores.
Both GPUs have 3 fans each and large heatsinks. Zotac’s version has its own Icestrom cooling technology, while Colorful uses an X radiator and X backboard system. During gaming, Zotak and Colorful both had an average temperature of 68 degrees and a maximum temperature of 82 degrees Celsius.
Zotac’s AMP AERO has come into the market with a very aggressive and cool design. Blue and Green colors are given as the color theme of this GPU. Both GPU vendors offer utility software for overclocking and other modifications. Overclocking and another tweaking can be done with Zotac’s GPU Firestorm and Colorful’s GPU iGame Center software.
Nvidia has fixed the price of this year’s RTX 4090 at 1600 USD, which is 100 USD more than the previous generation, 3090. Although the next generation has attractive upgrades like Articor and DLSS, they have not brought any changes in the memory section.