The new gen comes at an extra cost with more power and perfromance but do you really need it
WePC is reader-supported. When you buy through links on our site, we may earn an affiliate commission. Prices subject to change. Learn more
With the new RTX 4000 series of GPUs coming out, there are a lot of things to consider. Such as if it’s in fact worth upgrading or getting the new generation. So we consider the RTX 4090 vs 3090 Ti and if you can get a better deal from the last gen.
The Ada architecture debuts with the RTX 4090, which launched on the 12th of October 2022. Bringing plenty of improvements with it, such as raw FPS or tech such as DLSS 3. Along with improvements to current specs, it brings a large jump above what we may need right now.
Nvidia’s latest Blackwell GPU is set to go live today, below are the latest listings from the biggest retailers.
Prices and savings subject to change. Click through to get the current prices.
Whilst the RTX 3090 Ti was released on the 29th of March 2022. Coming after a load of delays and wonderment as to the reason it was coming. It further pushed what the Ampere spec could do and was seen as a test for the 40 series before it came out.
If you’re out looking to buy a 4090 you might be struggling you can check out our stock tracker or where to buy pages for help
Excellent air cooling solution
Pushes performance above the rest of the selection
Updated design cooling and aesthetically focused
A more expensive choice over the rest of the picks and will cost a lot to run
Excellent cooling
Potential for higher overclocking
Excellent performance
More expensive with AIO
Different series, and different architectures, mean a different way of working that makes it harder to compare directly. Between these, we see the change from Ada to Ampere architecture and a move in the foundry used for them. So we have a closer look at them.
The RTX 4090 moves back to TSMC and utilizes its 5nm process. In particular, it contains the AD102-300 variant. And with its build, the 608mm² die size fits in it 76.4 billion transistors. A huge increase over its previous sizing.
As the RTX 3090 TI is instead built with Samsungs 8nm process. The GA102-350 chip is 628mm² in size that is capable of holding 28.3 billion transistors. A much lower density over the newer card.
| RTX 4090 | RTX 3090 Ti | |
|---|---|---|
| GPU | AD102-300 | GA102-350 |
| CUDA Cores | 16,384 | 10,752 |
| RT Cores | 128 | 84 |
| TMUs | 512 | 336 |
| Memory | 24GB GDDR6X | 24GB GDDR6X |
| Base clock | 2,235 MHz | 1,560 MHz |
| Boost clock | 2,520 MHz | 1,860 MHz |
| Memory clock | 21.2 Gbps | 21 Gbps |
| Card bus | PCIe 4.0 x16 | PCIe 4.0 x16 |
| Memory bandwidth | 1,018 GB/s | 1,008 GB/s |
| TBP | 450 W | 450W |
Here we have a quick rundown of how the two cards differ. With that, we go through what that means for the two options.
Compute Unified Device Architecture are the cores in Nvidias GPU. These are equivalent to CPU ones and can work in unison to calculate the graphics for your display. These do change across the generations and so it’s not a fair comparison to look between them.
Even so, we see the count increase by 52% in the new flagship. Offering up some of where the big performance increase comes from in the new gen. All of those are capable of working together to create the graphics you see.
There are other available cores and we see those increase too. Such as the ray tracing cores, and TMUs that also help in particular tasks.
The frequency or clock rate is how fast these cores work through their instructions. As such a faster rate offers better performance. And between the generations, we see a big uplift. Likely coming from the more capable core and different transistors over the last gen.
The base clock itself jumps up 43%, whilst the boost clock is up 35%. With the custom cards likely offering better overclocks. So leading to a better performance of the cards.
The total board power is the amount of energy the card will consume at maximum utilization. It does offer a look at how much it is to run but also how much is need to be cooled. These two cards have the same 450W. But this time around there has been a lot more time to design coolers that can look much nicer and perform better. And the newer card is capable of doing more with the same amount.;
When it comes to the size of GPUs, the higher TBPs have led to a growth in their sizing. Both these cards have the same size, so you’d expect similar sizing. But the newer 4090 has had time to refine the cooling solutions.
So when it comes to the reference Founder’s Edition, the 4090 vs 3090 Ti size is in favor of the 4090. As it manages to shorten its dimensions with improved cooling. So it is down to 304 x 137 x 61mm compared to the 3090 Ti’s 336 x 140 x 61mm sizing.
But that doesn’t stop custom AIB partners from further pushing the sizing of the cards. We see that with the ASUS TUF for one. With the 4090 measuring in at a 4-slot 348.2 x 150 x 72.6 mm compared to the 3090 Ti’s 325.9 x 140.2 x 62.8 mm.
Overall you would expect to see a great increase in the performance of the cards with the new gen. And it’s not surprising that it does. When it came to the 3090 Ti vs 3090 it was a bit disappointing, now it’s quite a bit better.
As in 4k, we see the RTX 4090 improve by 55% over the 3090 Ti on average. And that is the story across the board. It has more power and tech available to it and so can achieve an 83% better in Blender as well. Whilst the ability to use DLSS 3 can push it into far beyond good performance even in the most intense situations.
Read our ROG Strix OC RTX 4090 review to see how it performs against a 3090.
Now in terms of pricing, there are some interesting factors. As the 3090 Ti launched for $1,999 USD. Whilst the 4090 comes in at only $1,599 USD/ £1,649. So it would be easy to say the 4090 is the better pick.
However, now pricing has been dropping for the 30 series. And so you can easily find the 3090 Ti at around $1,00 instead so even below the new card. So it can be a deciding factor.
Updated cooling to adapt to the higher power rating
Comes with a GPU support to help with the heavier build and prevent bending
Strong design with a light feature
A bit more of a dated design
High quality and design
Plenty of thermal capacity
Stylistic design
More expensive OC option
Outright, yes the RTX 4090 is much better than the 3090 Ti. It can perform much better with its higher specs and newer technologies. However, it comes in at a high price point and so it might be worth considering the older card, especially if you don’t need all the extra performance as the 3090 Ti still has a lot to offer.
You can also consider waiting to see what RDNA 3 has to offer. Coming out later it can offer a good alternative to the current launch.
Comparing the speed of these can show the advancement of the architectures and what they are capable of achieving. And the RTX 4000 series is a big increase in the clock speeds of the cards. And the RTX 4090 runs with a base clock of 2,235 MHz, and a boost clock of 2,520 MHz. Whilst the RTX 3090 Ti only runs with a base clock of 1,560 MHz and boosts to 1,860 MHz.
The RTX 4090 gets to 55% better in 4k performance over the 3090 Ti. Whilst only leading by 28% at 1440p and 15% at 1080p. However, the lead in ray tracing and rasterization is much improved as well.
Trusted Source
WePC’s mission is to be the most trusted site in tech. Our editorial content is 100% independent and we put every product we review through a rigorous testing process before telling you exactly what we think. We won’t recommend anything we wouldn’t use ourselves. Read more
