We can find graphics cards on the market for all tastes, needs and, of course, for all budgets. The manufacturer GALAX seems to have a new variant of the NVIDIA RTX 4070 ready, but it is not with a new heatsink. The new GALAX RTX 4070 has GDDR6 memorieswhen normally this model used GDDR6X type memory chips.
The most common changes in graphics cards are changes in the heatsink design and the use of different GPUs. It is quite normal to use different types of GPUs for the same model, but this is less common.
It is unusual for a model to receive a memory change and more, from a manufacturer. We do not know the reasons why GALAX (KFA2, in Europe) launches this model of graphics card. It could be focused on the Asian market, making it a little cheaper, but it is speculation.
GALAX launches an RTX 4070… with GDDR6 memory!
First of all, you should know that all RTX 4070 models (be it regular, Ti or SUPER) use GDDR6X memory. All of these memory modules have a memory speed of 21 Gbps. So this new model is a rather strange thing.
GALAX is currently finalizing the launch of a new RTX 4070 OC 2X with GDDR6 memory. The design of the graphics card and the box are exactly the same as the RTX 4070 model they already have on the market. This is a problem for customers, as they will have to pay close attention to the details of the packaging and the specifications on the website of the store where they buy.
It maintains the simple dual-fan design with double-slot thickness, so physically there are no changes. The single 8-pin power connector is also maintained, which is normal, since consumption hardly changes.
We don’t know if the PCB is still the same as the GDDR6X model, but that would be normal. This is more speculation, as there is no image of it to verify this.
For this card, the AD104-251 GPU will be used instead of the AD104-250 silicon. Actually, as far as the GPU is concerned, there are no changes in the configuration, maintaining the 5,888 CUDA Cores and other specifications. The only thing that changes in this card is the memory.
Specifically, it will use GDDR6 memory with a memory speed of 20 Gbps. Currently, the RTX 4070, in its ‘normal’, Ti and SUPER variants, use GDDR6X memory with a speed of 21 Gbps. The 12 GB capacity is maintained and the memory bus remains at 192 bits.
In principle, performance should be practically the same on both models. The GDDR6 variant may be 1-5% slower in games, a variation that is practically imperceptible. We do not know if this model will be cheaper, since no prices have been given.
We are confused about this graphics card, since NVIDIA does not offer this option. We have already seen similar things in the past, although it is usually with improved memory, not a worsening. We will see if it is launched on the market in the next few days and all our doubts can be resolved.