This is how the NVIDIA RTX 2070 Max-Q for laptops performs against the GTX 1070 Max-Q.
Contents
CES 2019 where NVIDIA presented its RTX Max-Q is far away, but it was relatively recently when Huang’s and the assemblers gave the go-ahead for their laptops equipped with these GPUs. Now, with enough options to choose from, it’s time to look at the scalability of the architecture against Pascals. How will the RTX 2070 Max-Q fare against the GTX 1070 Max-Q?
RTX 2070 Max-Q vs GTX 1070 Max-Q: Feature comparison doesn’t hold up
With the release of the first Max-Q under the Pascal architecture, NVIDIA achieved something unusual: introducing the same chip of a desktop GPU in a common laptop or a slim one.
The features were almost identical to the first ones, so laptop users could take advantage of the power offered on the desktop, but in a mobile way, justifying the price of these devices.
In the case of the RTX 2070 Max-Q we have to distinguish between its characteristics, since, although the GPU is physically the same and carries the 2304 shaders, its frequencies have been cut substantially.
Thus, the desktop RTX 2070 Founders Edition carries speeds of 1410 MHz on base and 1710 MHz under boost, while the Mobile version drops its frequencies to 1215/1440 MHz.
The Max-Q version sees the frequencies further cut, to 885/1185 MHz respectively. This should not be surprising, but what should be is the great gap that now exists between models.
Consumption is the main culprit for the different performance between models
As expected, the fault lies mainly with TSMC’s 12nm process, since NVIDIA cannot currently have TSMC’s 7nm and Samsung will go directly to EUV.
This produces, together with a larger die and millions of more transistors, a fairly high power consumption for laptops.
In the GTX 1070 we had 115 watts for the Mobile and Max-Q models, where we now start with the same consumption, but from the 175 watts that the desktop RTX 2070 carries.
This forces NVIDIA to lower the clocks more than they would like, limiting performance, where they have also had to include 12 Gbps GDDR6 to further reduce consumption and therefore performance even more.
Clock stability under stress
One of the key points in any gaming laptop is the cooling of the graphics card. It is the component that generates the most heat and therefore it is the focus that must be dissipated in the first place.
Each laptop model will differ in this regard, so it is important to take into account which model we choose.
As we can see, in the Gigabyte Aero 15 X9, its fan configured in game mode, the frequencies are always above the boost clock, although with less stability than what was seen in the GTX 1070 Max-Q, which does not have to affect the frametime in yes, as long as you keep the framerate stable.
RTX 2070 Max-Q vs GTX 1070 Max-Q Performance Summary
As we can see, except for Hitman, in all the games the new RTX manages to surpass the already veteran GTX.
The problem is that the total difference is only 10.6% over the performance offered by the Pascal-based card, which is initially a small difference for someone who can spend 2000-2500 euros on a laptop that carry this new Turing card.
If these narrow differences are maintained in higher models, which suffer from the same problems as this RTX 2070 Max-Q in terms of frequencies, we could be talking about a new NVIDIA fiasco, since contrary to what is thought, these versions they are ported to laptops that are usually more expensive than those that include regular Mobile versions.
