NVIDIA RTX 4090 FE:
The :
As a reminder, the cards of the ” Founders Edition ” series are unique models offered by NVIDIA. The main advantage of these cards is to be proposed at a “locked” price that will not change in the weeks or months to come. It will be available this October 12 at a price of $1 599. NVIDIA’s partners will also offer their own model, but at a higher price, since few of them manage to match the Founders Edition price. Visually, NVIDIA has kept a very similar design to the RTX 3090 Ti FE as well as its thickness of 3 slots.
As usual, NVIDIA has taken particular care with the packaging of its card by offering a cardboard box where the RTX 4090 FE is hidden. In the bundle, we find a 16-pin (12+4) to four 8-pin 150w adapter (600w in total). This adapter is necessary for those who do not have a power supply with a 16-pin port. Please note that NVIDIA is the only supplier of the adapter and therefore provides it to its partners. We have also received the 16-pin connector from Corsair, which is compatible with the high-end models of the brand such as the H1500i but also the AX1600i.
The NVIDIA RTX 4090 Founders Edition benefits from NVIDIA’s brand new architecture, namely “ADA Lovelace”. It replaces the Ampere architecture of the GeForce RTX 30xx launched a little over two years ago.
This flagship model of NVIDIA has a GPU AD102-300 with 16384 CUDA cores or 5632 cores more than the RTX 3090 Ti. It is equipped with a manufacturing process in 4nm from TSMC, not the 8 nm manufacturing process made by Samsung which allows to lower the TGP to 450 watts and not to 600 watts as we thought for several weeks. This change of engraving has had an impact on the size of the heat sinks as we discuss in this brief. The Boost frequency is also very different as it is 2520 MHz versus 1860 MHz on the RTX 3090 Ti FE.
In terms of memory, there are 24 GB of GDDR6X memory in 384-bit, which translates into a bandwidth of 1008 GB/s or 1 TB/s. The speed of this RTX 4090 should be 2 to 4 times faster than the RTX 3090 Ti according to NVIDIA. On the next page, we will come back in details on the new features brought by this new architecture, including the support of DLSS 3.
Cooling:
NVIDIA has kept the proven cooling principle of the previous generation. However, there is a difference in that the two fans are now 120 mm models with 7 blades that increase the airflow by 20%.
The goal is, as with any cooling system, to keep the GPU cool while producing acceptable noise levels. So NVIDIA had the idea to use two fans but placed on both sides of the graphics card. The one on the left sucks air from inside the case and is then expelled from the back of the card. The other fan, located on the right, also draws air from your case to dissipate the heat from the vapor chamber before being expelled in the upper position, still within your case. Needless to say, you’ll need to opt for good air circulation within your case if you want to keep your components cool.
The power stage consists of 23 phases and 12 2 GB GDDR6X memory chips. The RTX 4090 uses a PCIz 4.0 x16 interface but its power architecture is compliant with PCIe Gen 5 specifications. Finally, there is no RGB backlighting but just the GeForce RTX logo on the top edge that lights up in white.
