NVIDIA GeForce RTX 4090, RTX 4080, RTX 4070, RTX 4060, RTX 4050 – specification of Ada Lovelace graphics systems for laptops

CES this year is apparently the perfect opportunity to introduce a new generation of graphics chips and processors for laptops. It’s been a long time since you could receive top graphics chips as well as processors from Intel and AMD at a single fair. While AMD will present its debut APUs Pheonix and Dragon Range only tomorrow, Intel and NVIDIA do not hesitate and reveal the details today. We have already written about Raptor Lake processors, now it’s time to present the new generation of NVIDIA GeForce RTX 4000 Laptop GPU mobile graphics systems. There will be some surprises in the specification.

NVIDIA officially introduces the next generation of GeForce RTX 4000 Laptop GPUs, based on the Ada Lovelace architecture.

Intel Raptor Lake-HX, Raptor Lake-H, Raptor Lake-P and Raptor Lake-U – premiere of 13th generation processors for laptops

The specification of the most powerful NVIDIA GeForce RTX 4090 Laptop GPU and GeForce RTX 4080 Laptop GPU systems is consistent with what was previously reported on the network, at least as far as Ada Lovelace core configurations are concerned. The top new generation system uses the AD103 core with 9728 CUDA cores and 16 GB of GDDR6 memory on a 256-bit bus. Unfortunately, NVIDIA did not provide us with information about the speed of the GDDR6 bones, and therefore the overall VRAM bandwidth. In the case of the GeForce RTX 4080 Laptop GPU, we will get an AD104 core with 7424 CUDA cores and 12 GB of GDDR6 memory on a 192-bit bus. Both systems have a default TGP of a minimum of 80W to a maximum of 150W, with the possibility of increasing an additional 25W thanks to Dynamic Boost technology (from 80 to 105W and from 150 to 175W respectively).

NVIDIA GeForce RTX 4090 Laptop GPU will receive AD103 chip with 9728 CUDA cores with performance above GeForce RTX 3090

The next two systems are already quite strong castrates and here I must admit that I am a bit worried about the real performance in games. The NVIDIA GeForce RTX 4070 Laptop GPU is based on the AD106 core and has 4608 CUDA cores on board. In turn, the GeForce RTX 4060 Laptop GPU uses the AD107 core with 3072 CUDA cores. A short comparison is enough to see that both systems from the Ada Lovelace family have fewer computing units than their theoretical predecessors. Theoretical, because if the nomenclature was kept more unified with desktops, the former were in fact the GeForce RTX 4060, and the latter at most the RTX 4050 Ti system. Both systems also work on a 128-bit bus and have 8 GB of GDDR6 memory. While in the case of the GeForce RTX 4060 we get (finally!) 2 GB of memory more than its predecessor, leaving 8 GB for the GeForce RTX 4070 is a kind of not funny joke. At the end there is the NVIDIA GeForce RTX 4050 Laptop GPU with a truncated AD107 core (2560 CUDA cores, so slightly less than the RTX 4060) and 6 GB of GDDR6 memory on a 96-bit bus. As a reminder, the predecessor offered the GA107 system with 2048 CUDA cores and 4 GB of GDDR6 memory on a 128-bit bus. All Ada Lovelace mobile systems support NVIDIA DLSS 3 technology, and the two most powerful versions also have the new Dual AV1 codec, known from the desktop GeForce RTX 4000 models.

As for the TGP of the other Ada Lovelace graphics systems, there are also changes compared to the previous generation of Ampere. The GeForce RTX 4070 Laptop GPU, GeForce RTX 4060 Laptop GPU, and GeForce RTX 4050 Laptop GPU have a much wider range – from just 35W to 115W, plus the ability to increase by 25W with Dynamic Boost. Thus, the most energy-efficient version of these systems can work with 35 W, and the most packed with 140 W. The biggest differences can be seen in the context of the GeForce RTX 4050 Laptop GPU, where the predecessor offered a maximum of 95 W with Dynamic Boost. Now at 140W, the performance differences in this GPU class can be much more noticeable.

Source: PurePC