NVIDIA May Downsize GeForce RTX 5090 Founders Edition Card
TapTechNews May 30th news, the source @kopite7kimi yesterday retweeted the overclock3d media report, revealing that NVIDIA is considering downsizing the GeForce RTX 5090 Founders Edition graphics card size to a dual-slot dual-fan design.
The current NVIDIA GeForce RTX 4090 and RTX 4080 Founders Edition GPUs use a three-slot and dual-axis technology fan design. If the news is true, it is not clear what design plan the GeForce RTX 5090 Founders Edition will adopt.
There are two possibilities:
A dual-slot design means that NVIDIA has completely changed the cooling solution to allow it to dissipate more than 400 watts in a smaller package.
NVIDIA's Blackwell GPU is a very excellent chip that can achieve high performance without generating a lot of heat.
TapTechNews quoted RedGamingTech's revelation in February this year. The RTX 5090 graphics card is equipped with 192 streaming multiprocessors (SM), and will exceed the Lovelace flagship in terms of architecture. In addition, compared with GB102 and AD102, the rasterization performance has increased by 60%.
RedGamingTech also said that in terms of ray tracing, the performance target of NVIDIA RTX 5090 is to increase by 2.5 times, and the computing power will increase by 2 times, and the power consumption may be as high as 520W.
And recent related news indicates that the RTX 5090 FE public version adopts a three-PCB design and leaves space to achieve double-sided ventilation.
The memory bit width is 512-bit and requires 16 GDDR7 memory chips, and the memory capacity reaches 32GB. With the improvement of the memory configuration, the layout of the memory chips will change from the 3-4-1-4 of the RTX 4090 to 4-5-2-5 (clockwise direction), that is, 4 on the top, 5 on each side, and 2 on the bottom, a total of up to 16.
