At GTC 2022, Nvidia unveiled the first lineup in the RTX 40-series of GPUs. It follows after months of speculation after a series of leaks.
The 4090 will be available for $1599 and releases on October 12th. It is two times faster than the 3090 Ti in Microsoft Flight Sim. Alongside this, the 4080 was announced. While it doesn’t have a date tied to it, it will launch at two prices, due to there being two versions: 4080 16GB and 12GB.
The two prices currently being presented by Nvidia are $1199 for the 16GB version, while the 12GB is expected to launch at around $899.
Ada Lovelace is the name of the new RTX 40-series cards, using 76 billion transistors. Nvidia has worked with TSMC to produce a 4nm process to build the new Ada Lovelace cards.
Alongside the card, Nvidia has also announced that DLSS 3 will be coming, which they say will bring further improvements to the already impressive DLSS 2. It works off of a 16K reference image to train the deep learning engine behind it.
DLSS 3 introduces new ways to process the image, looking at the upcoming and previous frames. It generates a new frame, and DLSS 3 takes the work off of the GPU and CPU. This allows them to hit a four times performance boost over brute-forced rendering.
In a demo, they showed off Cyberpunk running at 96 FPS with all the bells and whistles running.
In Flight Simulator, DLSS 3 allows GPUs and CPUs to run the game with up to double the performance.
Nvidia also announced Portal RTX, a new version of the game that is in similar scope to Quake 2 RTX. A complete retexturing of the game to take advantage of raytracing and DLSS 3. It’ll be a free piece of DLC for owners of the game.
They have also introduced a new mode to Omniverse, to allow modifications to older games, allowing keen fans and developers to update their old games with new textures. It is called RTX Remix.
RTX 4090 and 4080 12GB and 16GB specs:
|Spec||RTX 4090||RTX 4080 16GB||RTX 4080 12GB|
|NVIDIA CUDA Cores||16384||9728||7680|
|Boost Clock (GHz)||2.52||2.51||2.61|
|Base Clock (GHz)||2.23||2.21||2.31|
|Standard Memory Config||24 GB GDDR6X||16 GB GDDR6X||12 GB GDDR6X|
|Memory Interface Width||384-bit||256-bit||192-bit|
|Maximum GPU Temperature (in C)||90||90||90|
|Graphics Card Power (W)||450||320||285|
|Required System Power (W) (4)||850||750||700|
|Required Power Connectors||3x PCIe 8-pin cables (adapter in box) OR1x 450 W or greater PCIe Gen 5 cable||3x PCIe 8-pin cables (adapter in box) OR1x 450 W or greater PCIe Gen 5 cable||2x PCIe 8-pin cables (adapter in box) OR1x 300 W or greater PCIe Gen 5 cable|
|Maximum Digital Resolution (1)||7680×4320||7680×4320||7680×4320|
|Standard Display Connectors||HDMI(2), 3x DisplayPort(3)||HDMI(2), 3x DisplayPort(3)||HDMI(2), 3x DisplayPort(3)|
|Length||304 mm||304 mm||Varies by manufacturer|
|Width||137 mm||137 mm||Varies by manufacturer|
|Slot||3-Slot (61mm)||3-Slot (61mm)||Varies by manufacturer|
Jensen Huang’s keynote at GTC 2022 revolved around what Nvidia is bringing next to the world of computing. Their next lineup of graphics cards, the RTX 4000 series, has been announced. It follows after a brief tease and months of speculation.
GTC is Nvidia’s AI and technology conference, using the tagline ‘Project Beyond’ during the build-up in the last couple of weeks. The conference not only focuses on their hardware but also on the company’s investment in AI, the metaverse, and content creation. Alongside the announcement of the RTX 4090 and 4080, the keynote went onto talk about Nvidia’s advancements in corporate tech.
However, it doesn’t look like EVGA, a long-time partner of Nvidia, will be joining them on the RTX 40-series journey. The company has recently taken its business elsewhere after a dispute between themselves and Nvidia.