Nvidia’s new 4000 super lineup has launched. But what is so SUPER about these new GPUs? Keep reading.
Starting with the RTX 4070 super which has been priced almost the same as the previous non-super variant, has got a 20% increase in the core counts (which includes CUDA, TENSOR, and RT cores) and an increase in L2 cache by 30%. Just like how it was done with the 4000 series which translates to 10-15% increase in RAW gaming performance MAKING THE 4070’s PERFORMANCE EQUAL TO THE 3090.
And when we say 3090, we are talking about the two years ago’s best 4K gaming GPU costing around 1.5L which is now just around 70K for RTX 4070 Super.
Talking about the performance, it is a 10% increase but there is a catch! In 3D editing apps, due to VRAM still being at 12GB, it shows a dramatic dip as you can see in the benchmarks below.
But hey my friend, don’t feel sad because the 4070 Ti Super is here to fill the emptiness in the 4070 Super.
With increase in VRAM to 16GB it pushes performance and stability in 3D rendering apps and which in-turns make it an amazing option for 3D Creators.
Although on paper, it hasn’t received as much of a change except the core count increasing to 10% more than the non-super variant.
Which summarises to a 10-15% increase in gaming and content creation or 3D modelling compared to the regular 4070 Ti and also a very close gap between the 4070 Ti Super & the 4080.
So if you were looking to get a card that has 16GB of VRAM, now it’s possible with the 4070 Ti.
And speaking of AI (just like Nvidia), let’s talk about the 4080 super and where it really shines.
We know every tech-youtuber on the internet is bashing the 4080 for not giving a major uplift in performance as it is as low as 3% but it doesn’t make it a bad card.
Because firstly it’s ₹30,000 cheaper now, secondly it’s not even made for Gaming.
Yes! The future is AI so why not let’s start with new AI GPUs.
The 4080 super, as can be seen in benchmarks below, is 12% better in AI, even though the specs are just 7% better compared to the regular 4080. We’re not even sure how that’s possible but if you look at the benchmarks of all the new super GPUs – the biggest performance uplift is in AI Training & Inference Benchmarks.
4080 super ai benchmarks
Now, it’s no secret that 2024 is the year for AI. So as the start of Nvidia 40 Super Series GPUs.
But of course not everyone is buying a GPU for AI – so which will be the best option for you?
To summarise, for gaming or Adobe suite tasks, the RTX 4070 could be your choice as it supports 4K gaming, and since Adobe relies more on CPU, the 4070 provides sufficient software acceleration without the need for more.
And for the 3D Artists, we feel the 4070 Ti Super is the best choice. Like what we have said earlier in this post, the VRAM will come in really handy for handling large & environmental projects while costing ₹20,000 less than the 4080 Super.
And finally AI developers and people who want the best of the best, can look at the flagship, which is the 4080 super.
So that was our take on the Super Series. Do get in touch if you want us to build a PC for you with RTX 4000 Supers.
