NVIDIA GeForce RTX 4090 Laptop GPU vs NVIDIA RTX 3000 Ada Generation Laptop GPU

NVIDIA GeForce RTX 4090 Laptop GPU

The NVIDIA GeForce RTX 4090 Laptop GPU (Codename GN21-X11) is a high-end laptop GPU. It is based on the AD103 chip as the desktop RTX 4080 and use the Ada Lovelace architecture. The AD103 chip offers a maximum of 10.752 shaders from which inly 9728 are used (e.g., the RTX 4080 Desktop uses 9.728) and a 256 Bit memory bus. The RTX 4090 mobile uses 16 GB GDDR6 dedicated graphics memory with a clock speed of 20 Gbps (effective). The TGP (Total Graphics Power) can be configured between 80 and 150 Watt (+ 15 Watt Dynamic Boost from the CPU) leading to different performance levels (the desktop RTX 4080 uses 320 Watt for comparisons). 

The performance depenends on the TGP settings which defines the maximum boost clock (1455 MHz for 80W and up to 2040 MHz for 150W). With 150W the performance should be in the region of the desktop RTX 4080 and therefore best suited for 4k60 or 2k with 120 fps for demanding games. For raytracing, lower resolutions or DLSS is needed.

In our reviews, an average RTX 4090 is clearly the fastest mobile GPU and bests the RTX 4080 Laptop GPU and mobile RTX 3080 Ti easily. 4090M GPUs with lower power limit can however loose up to 20% performance.

In games, fast RTX 4090 models allow gaming in 4k with maximum details. Adding Raytracing may lead to frame rates below 60 fps (e.g. F1 23) and therefore the use of DLSS or reduced settings are advised.

The AD103 chip of the RTX 4090 mobile is manufactured in an improved 5 nm process at TSMC (4N).

NVIDIA RTX 3000 Ada Generation Laptop GPU

The Nvidia RTX 3000 Ada Generation, not to be confused with the A3000, P3000 or RTX 3000 Turing Generation, is a higher-end professional graphics card for use in laptops that sports 4,608 CUDA cores and 8 GB of ECC GDDR6 VRAM (error correction can be turned off if desired). It would be fair to say that this is a GeForce RTX 4070 Laptop in disguise; consequently, the graphics card is fast enough for many games at 1440p with quality set to High. Brought into existence in 2023, the RTX 3000 leverages TSMC's 5 nm process and Nvidia's Ada Lovelace architecture to achieve very decent performance combined with moderate power consumption. The Nvidia-recommended TGP range for the card is very wide at 35 W to 140 W leading to bizarre performance differences between different systems powered by what is supposed to be the same product.

Quadro series graphics cards ship with much different BIOS and drivers than GeForce cards and are targeted at professional use rather than gaming. Commercial product design, large-scale calculations, simulation, data mining, 24 x 7 operation, certified drivers - if any of this sounds familiar, then a Quadro card will make you happy.

Architecture and Features

Ada Lovelace brings a range of improvements over older graphics cards utilizing the outgoing Ampere architecture. It's not just a better manufacturing process and a higher number of CUDA cores that we have here (up to 16,384 versus 10,752); under-the-hood refinements are plentiful, including an immensely larger L2 cache, an optimized ray tracing routine (a smarter technique is used to determine what is transparent and what isn't), and other changes. Naturally, these graphics cards can both encode and decode some of the most widely used video codecs, AVC, HEVC and AV1 included; they also support a host of Nvidia technologies, including Optimus and DLSS 3, and they can certainly be used for various AI tasks.

The RTX 3000 Ada features 36 RT cores of the 3rd generation, 144 Tensor cores of the 4th generation and 4,608 CUDA cores, making it ~11% inferior to the RTX 3500 Ada Generation as long as we disregard clock speed differences. (The 3500 also has more VRAM). The RTX 3000 comes with 8 GB of 128-bit wide ECC GDDR6 memory for throughput of ~256 GB/s. The fact that error correction is present here proves that the graphics adapter is indeed targeted at professional users, and it can be disabled if desired.

Just like Ampere-based cards, the RTX 3000 makes use of the PCI-Express 4 protocol. 8K SUHD monitors are supported, however, DP 1.4a video outputs can potentially prove to be a bottleneck down the line.

Performance

Having reviewed the Dell Precision 5480 and its 50 W RTX 3000 Ada in September 2023, we can firmly say the graphics card is in the same league as the RTX 4060 Laptop and the RTX 4050 Laptop, despite sharing most of its specifications with the RTX 4070 Laptop. 48 seconds in the Blender 3.3 Classroom CUDA test and a 3DMark 11 Performance GPU score of less than 30,000 points are nothing to write home about. The GPU's gaming talents are not mind-blowing, either, although it is certainly capable enough for 1080p gaming.

• Tiny Tina's Wonderlands - almost 90 fps at 1080p / High
• Cyberpunk 2077 - over 60 fps at 1080p / High

Nvidia's marketing materials mention "up to 20 TFLOPS" of performance, a very noticeable improvement over 14.5 TFLOPS delivered by the RTX 2000 Ada Generation. Your mileage may vary depending on how competent the cooling solution of your laptop is and how high the TGP power target of the graphics card is. A 100 W RTX 3000 Ada might be able to come close to the average RTX 4070 Laptop.

One other thing worth mentioning is that enabling error correction appears to reduce the amount of video memory that is available to applications and games by up to a gigabyte.

Power consumption

Nvidia no longer divides its laptop graphics cards into Max-Q and non-max-Q models. Instead, laptop makers are free to set the TGP according to their needs, and the range can sometimes be shockingly wide. This is especially the case for the RTX 3000, as the lowest value recommended for it sits at just 35 W while the highest is 300% higher at 140 W (this most likely includes Dynamic Boost). This is the kind of delta that we've be seeing on consumer-grade laptops featuring the latest GeForce RTX cards.

Last but not the least, the improved 5 nm process (TSMC 4N) the RTX 3000 Ada is built with makes for very decent energy efficiency, as of mid 2023.