Hot Potato: If we think of the latest drop in leaks, gamers who thought existing RDNA2 and RTX 30 GPUs were currently in demand might be surprised.
In addition to pushing the GPU core as far as possible, the deal is provided Excellent performance for this budget is made possible by the constant use of hot and thirsty GDDR6X memory on Nvidia's top models.
450-500w-Greymon55 (@greymon55) Jul 29, 2021
Separately, Bondrewd of Beyond3D Forums references Navi 31, the top SKU of AMD's RDNA3 line, this indicates that The multi-chip GPU has a total board power of less than 500W and less than 350mm2 in each graphics core format. Approximately 600-650mm2 for just two GCDs and possibly 800mm2 for the whole GPU (when using Infinity Cache MCDs), 3Dcenter thinks the Navi 31 falls in the -450.480W range for total powerband. In addition to putting more pressure on PSUs and cooling in smaller cases, the increased power of GPUs is increasingly outpacing gaming laptops.
Restrictions on the size and weight of gaming laptops limit their ability to cool, and as a result TDPs have proven stubborn since mobile GPUs rarely run more than 150W And these larger models have a lot more cooling.
That was good enough: 180W GTX 1080 graphics were a flaw, but the RTX 2080 asked for 215W and the RTX 3080320W, gaming laptops in several generations are about half the power of a desktop card wins .
That doesn't matter the size of the GPU molding at the top of the laptop is completely enclosed, while the RTX 3080 and RX 6800M are actually a layer of silicon lower than desktop use. Peer-to-peer or partner cards typically draw more power reference numbers, such as the FTW3 EVGA, increasing the RTX 3080 to roughly 400 watts.
Sure, 4,000-watt GPUs aren't likely to be able to make premium fighters a problem with their unique cooling and kilowatt design. But for casual gamers - those who use small bags or laptops or use a reliable 500W PSU - they become a huge problem.
Next-gen GPUs look big and hungry, and that's bad news