Bringing out a new generation of products to replace others should be an obvious leap in performance. Especially if we talk about graphics cards where the important thing is that they draw the largest number of pixels on the screen and at the highest possible speed. Of course, the problem comes when external elements end up being limiting factors. Is it time for NVIDIA and trick AMD to start considering stopping scaling down their GPUs for laptop models?
The RTX 4070 Mobile is not the leap that was expected
On the “Jarrod’s Tech” YouTube channel they have decided to compare two practically identical laptops, but with the difference that one is equipped with an RTX 4070 Mobile and the other with the already veteran RTX 3070 Ti Mobile. The objective? Check the generational jump of NVIDIA graphics cards from one generation to another and it seems that the jump between the two is not what was expected. What’s more, the conclusion they have reached after testing several games at different resolutions is that both graphics cards are practically identical in terms of the power they offer in games.
Actually, the moment in which the RTX 4070 Mobile takes advantage is by applying the interpolation of frames to generate new ones, a functionality that is currently exclusive to the RTX 40, but from here we believe that if those of Jen Hsen Huang have decided to make it exclusive to the new generation. However, this is not the only limiting point we see, but the fact that being limited in terms of bandwidth in notebook VRAM is the biggest limiting factor.
Let’s not forget that NVIDIA can use GDDR6X memory in its GPUs for conventional PCs, in laptops, not because such video memory is unfeasible due to consumer issues. In other words, we have a GDDR6 bottleneck that prevents gaming laptops from scaling further.
Are they a bad buy for the end user?
On the contrary, they are an excellent purchase for gaming performance if you are going to play in Full HD and even under Quad HD, since the power of these graphics cards far exceeds what the games on the market need at the today to work properly and without performance problems. So the topic is not at this point.
The reality is that the RTX 3070 Ti Mobile is no longer manufactured and any product with said card will cease to exist in a short time within the models available on the market. Rather, the problem comes with the upward price trend that is taking over the entire general sector and that affects our pockets as a side effect of the crisis. Since we see prices rise, an increase in benefits is expected in accordance with said increase.
A harbinger of what the mid-range of NVIDIA’s RTX 40 will be like?
It’s no secret to anyone that laptop graphics cards are limited in terms of power, which means that different models have to scale their clock speeds down and that is precisely the problem that NVIDIA’s RTX 40 is having. for laptops. Let’s not forget that the desktop versions are based on taking advantage of the jump from 8 nm to 4 nm to be able to increase the clock speed and compensate for the fact that, except for the AD102 chip that we will never see in laptops, cutting the number of cores in order to have smaller chips.
And what is this due to? At the cost per wafer, that is, NVIDIA has decided not to raise it and hence if we look at the rest of the RTX 40 chips, that is, from AD103 onwards, we will see that they do not have more cores than the RTX 30 and with a bus more narrow. The way to alleviate it? Put a huge L2 cache pit to reduce dependency on external memory. Despite the cuts, the use of the new connector allows you to increase the clock speed and completely alleviate these cuts. The problem comes when we scale within the same generation to lower and lower consumption powers and where the high speeds in GHz can no longer be as high. That’s when you see the limitations.
