NVIDIA just took the wraps off their new RTX 40 series Geforce and RTX 6000 workstation GPUs during their GTC 2022 keynote. The updated GPUs bring massive performance gains to the table alongside the new Ada Lovelace architecture, TSMC 4N process, and much more.
For Geforce, the company has mainly focused on improving Ray Tracing capabilities and bumping the Gaming performance significantly. The biggest improvement however comes in the form of DLSS 3 which can generate entire frames in games using AI (Deep Learning). This will put the GPU in the driver’s seat and remove the limitations of the CPU to allow for faster and more intelligent gameplay.
The RTX 6000 is going to appeal more to professionals in the graphics industry who need as much power as they can get. The GPU offers massive improvements all across the board over its predecessor.
NVIDIA also has improved, dual NVENC Media Encoders that now support AV1 video and this also means a massive boost in intensive video workflows. It’ll also allow users to record 8K 60FPS gaming videos in real-time, which is pretty insane.
Table of Contents
NVIDIA Geforce RTX 40 Series
We got two new Geforce GPUs this time around including the flagship RTX 4090 and the RTX 4080. The RTX 4090 comes in a single configuration while the RTX 4080 has two variants. The entire line-up has received all the generational improvements and you’ll be paying a premium if you want one of these new GPUs.
Geforce RTX 4090
The NVIDIA RTX 4090 is the most powerful GPU announced by the company at this year’s GTC. It is a pretty nice bump over the RTX 3090 Ti from last year and this gap is much more significant when compared to the regular 3090. However, we didn’t get a memory bump as NVIDIA stuck with 24GB GDDR6X. Impressively enough, power consumption is also the same at 450W.
For games, the company claims that the RTX 4090 is up to 4 times faster using DLSS 3 compared to the 3090 Ti which uses DLSS 2. In general, for games that can’t embrace it all, the bump is more like 2 times. For video editors, the improvements should also be a lot, but we’ll need to see the real-world performance before we can confirm any of these.
RTX 4090 Specifications
- Boost Clock: 2.52 GHz
- Base Clock: 2.23 GHz
- Architecture: Ada Lovelace
- Memory: 24GB GDDR6X
- Memory Interface Width: 384-bit
- CUDA Cores: 16384
- RT Cores Type: 3rd Gen
- Tensor Cores Type: 4th Gen
- Maximum Display Resolution: 7680 x 4320 (8K)
- HDCP: 2.3
- HDMI: HDMI 2.1a
- Display Port: 1.4a
- Power: 450W
- Dimensions: 304 mm (L) x 137 mm (W)
The Geforce RTX 4090 will cost you $1599 and it’ll be available starting October 12, 2022. Yes, this is a lot of money for a Geforce series GPU and if you don’t need this much power, you can go for the RTX 4080 instead.
Also Read: The Dual-Screen Lenovo ThinkBook Plus Gen 3 Is Now Available In The US
Geforce RTX 4080
The more mainstream RTX 4080 has two variants this time around including one with 16GB GDDR6X memory and one with 12GB GDDR6X memory. However, although NVIDIA would like you to believe that these are the same GPUs, the two variants have multiple differences and it feels like the company didn’t name the 12 Gig variant the 4070 Ti just to put a slightly higher price tag on it.
Anyway, the RTX 4080 (both variants) did receive the same generational improvements as their more expensive sibling. So, you get DLSS 3, so modern games will be able to simulate frames and that makes it about 2 times faster compared to the RTX 3080 Ti.
The performance improvements put these far ahead of their predecessors and push them somewhere close to the RTX 3090 Ti and RTX 3090 territory, but do remember that you will see fewer gains if the game hasn’t adapted the 3rd gen DLSS.
RTX 4080 16GB Specifications
- Boost Clock: 2.51 GHz
- Base Clock: 2.21 GHz
- Architecture: Ada Lovelace
- Memory: 16GB GDDR6X
- Memory Interface Width: 256-bit
- CUDA Cores: 9728
- RT Cores Type: 3rd Gen
- Tensor Cores Type: 4th Gen
- Maximum Display Resolution: 7680 x 4320 (8K)
- HDCP: 2.3
- HDMI: HDMI 2.1a
- Display Port: 1.4a
- Power: 320W
- Dimensions: 304 mm (L) x 137 mm (W)
RTX 4080 12GB Specifications
- Boost Clock: 2.61 GHz
- Base Clock: 2.31 GHz
- Architecture: Ada Lovelace
- Memory: 12GB GDDR6X
- Memory Interface Width: 192-bit
- CUDA Cores: 7680
- RT Cores Type: 3rd Gen
- Tensor Cores Type: 4th Gen
- Maximum Display Resolution: 7680 x 4320 (8K)
- HDCP: 2.3
- HDMI: HDMI 2.1a
- Display Port: 1.4a
- Power: 285W
- Dimensions: N/A
As you can see, the 12GB variant of the RTX 4080 is technically a completely different GPU. If you take the 4090 and scale down some of the specs, you get the 4080 16GB variant and if you do that again, you get the 4080 12GB variant.
But the company decided to call both variants 4080 and you’ll be paying $1199 for the higher-end 16GB variant while the base 12GB one is $899. If you want to pick one up, these will be available in sometime in November 2022.
NVIDIA RTX 6000
In the workstation category, we got the RTX 6000 which also features the new Ada Lovelace architecture and noticeable improvements over its predecessors. Bringing more power to the creative lineup, the new 6000 brings better rendering capabilities, AV1 encoding & decoding support, better power efficiency, and more.
The major upgrades come in the form of a larger L2 Cache, an increased number of next-gen performance cores with a significant performance bump, and increased memory bandwidth compared to the previous generation. You’ll likely see massive gains in graphics rendering and AI training with the RTX 6000.
RTX 6000 Specifications
- Architecture: Ada Lovelace
- Memory: 48GB GDDR6
- Maximum Display Resolution: 7680 x 4320 (8K)
- Power: 300W
- Connector: 4x Display Port 1.4a
- Dimensions: 266.7 mm (L) x 111.7 mm (W)
There is no pricing info right now, but the NVIDIA RTX 6000 GPU will be available starting December 2022.
Also Read: Apple’s M3 and A17 SOCs Will Use TSMCs Upgraded 3nm Process Technology
RTX 40 Series and RTX 6000 – My Takeaway
The new Ada Lovelace architecture-based RTX 40 series and RTX 6000 series GPUs are impressive and they bring some much-needed improvements to PC graphics. However, if you’re already using a Geforce RTX 30 series GPU on your Desktop, I see no reason to rush out and upgrade. The same goes for those with laptops equipped with 30 series mobile GPUs. Looking at the desktop GPU line-up, the mobile variants are expected to perform pretty similarly.
However, if you’re coming from the RTX 20 series, the new RTX 40 series truly brings massive gains. So, for both 20-series Desktop and Laptop users, the 40 series is a worthy upgrade. The RTX 6000 brings some real improvements over the previous generation, so although you don’t necessarily need to upgrade, if you’re buying a new workstation GPU, invest in the RTX 6000 instead.
