GPU Pulls up to 200W, GA103-based, Lineup Power Detailed for NVIDIA GeForce RTX 4090 Laptop GPU
NVIDIA’s 2023 International CES event is anticipated to see the introduction of its GeForce RTX 40-series Laptop GPU series, which will power the next-generation gaming notebooks based on the upcoming 13th Gen Core “Raptor Lake” processors, as well as its desktop GeForce RTX 4070 and RTX 4070 Ti graphics cards. In this power- and thermally limited form-factor, NVIDIA appears to be walking a very fine line between power management and generational performance gain. On the specifications of several RTX 40-series Laptop GPUs, Wccftech earned a significant scoop.
The RTX 4090 Laptop GPU, based on the 4 nm “AD103” silicon, will be the highest model in the GeForce RTX 40-series “Ada” Laptop GPU portfolio (same one that powers the desktop RTX 4080). It will have 16 GB of memory, an unspecified number of cores, a GPU Boost frequency of up to 2.04 GHz, and a typical power draw of 150 W to 175 W with a 25 W dynamic boost range, with a peak power of 200 W (power permitted by the platform if the other components, like the CPU, aren’t drawing their peak power).
As we move down the list, we can see that the RTX 4080 Laptop GPU is based on the same “AD104” architecture as the future RTX 4070 Ti and RTX 4070 desktop graphics cards. Although the core configurations are yet unclear, judging by the name, the RTX 4080 Laptop GPU probably pushes it to the limit. The SKU includes 12 GB of memory spread across the 192-bit memory bus of the chip, boosts to 2.28 GHz, and has a comparable power-limit to the RTX 4090 Laptop GPU—150 W to 175 W normal graphics power, with additional leeway for 25 W.
Based on the as-yet-unreleased 4 nm “AD106” technology, the RTX 4070 Laptop GPU has 8 GB of memory and, if we’re not mistaken, has a 128-bit GDDR6X memory interface. With provision for an additional 25 W, this GPU receives standard graphics power in the 115 to 140 W range. The RTX 4060 Laptop GPU and the RTX 4050 Laptop GPU, which are located lower on the stack, are both based on the 4 nm “AD107” silicon, which is arguably the smallest “Ada” implementation. Memory for the former is 8 GB, while memory for the latter is 6 GB.
