NVIDIA GeForce 940MX vs NVIDIA GeForce RTX 4070 Laptop GPU

NVIDIA GeForce 940MX

The Nvidia GeForce 940MX is a mid-range laptop graphics card with DirectX 12 support (FL 11_0). It has been announced in the first quarter of 2016 and is the refresh of the GeForce 940M. Compared to the older card, the new one is based on the same chip but now supports faster GDDR5. Therefore, it is similar to the Maxwell GPU in the Microsoft Surface Book but with more memory and higher clock speeds. However, most notebook vendors will use the slower DDR3 version of the 940MX that clocks only a few percent faster than the old GeForce 940M (1122 versus 1072 MHz base clock e.g.). 

For the GDDR5 version of the 940MX there appears to be also a variant with the GM107-B chip with 512 shaders instead of the 384 but lower clock speeds. The 940MX in the Acer Aspire E5-774G e.g. featured 512 shader cores (GM107) clocked at 795 - 861 MHz. The Dell Inspiron 17 7778 version however had the GM108 chip with 384 shaders clocked at 1083 - 1189 MHz.

The performance greatly depends on the used graphics memory and chip. According to our benchmarks (see below), the DDR3 version is very similar to the old 940M (due to only minor increased clock rates). The GDDR5 version however pushes ahead 20% (even 30% if paired with a fast desktop quad-core, but we lack 940M reference scores). Therefore, demanding games from 2015 should run fluently in low to medium settings and WXGA resolution. 

The TDP of the card should be rated at about 15 to 30 Watt (depending on the version, clock speeds and memory). Therefore, the GPU is best suited for laptops from 13-inch and up.

NVIDIA GeForce RTX 4070 Laptop GPU

The NVIDIA GeForce RTX 4070 Laptop GPU (Codename GN21-X6) is a high-end laptop GPU. It is based on the AD106 chip and use the Ada Lovelace architecture. The RTX 4070 offers 4,608 cores (of the 7,680 in the AD106) and a 128 Bit memory bus. The RTX 4070 mobile uses 8 GB GDDR6 dedicated graphics memory with a clock speed of 16 Gbps (effective). The TGP (Total Graphics Power) can be configured between 35 and 115 Watt (+ 15-25 Watt Dynamic Boost from the CPU) leading to different performance levels (the desktop RTX 4070 Ti uses 285 Watt for comparisons). The boost clock speed ranges from 1230 MHz (at 35 Watt TGP) up to 2175 MHz (115W). Therefore, the performance is also depending a lot on the power setting used for the RTX 4080 mobile in the laptop.

The performance should be below the desktop RTX 4070 Ti and therefore best suited for 2k and maximum details for demanding games of 2022.

Our synthetic benchmarks of an average RTX 4070 mobile show a performance slightly below a mobile RTX 3080 Ti. However, with reduced power limits, the performance is clearly suffering, e.g. only 70% in Time Spy for a version with 60W TDP. In games, a fast RTX 4070 is best suited for QHD gaming, where some very demanding games wont reach the 60 fps limit (and may have to run at 1080p, reduced settings or using DLSS). See below for detailed gaming benchmarks of various cards.

The RTX 4070 offers 144 tensor cores (4th gen) that enable DLSS 3 including frame generation for supported games. Furthermore, 36 raytracing cores (4rd gen) can be used for raytracing effects in supported games.

The AD106 chip is manufactured in 5nm at TSMC (4N process).