Nvidia RTX 4090 vs Nvidia RTX 4080: Which Lovelace GPU is better?
Nvidia recently revealed its latest batch of GPUs, code-named Ada Lovelace. But how do the latest releases compare to each other?
During the Nvidia GTC 2022 Keynote, the company revealed two new GPUs in the RTX 4000 Series: the Nvidia RTX 4080 and the Nvidia RTX 4090.
But how do these two new GPUs actually compare, and what does the flagship RTX 4090 have that the RTX 4080 doesn’t? Read on to find the key differences between the hardware and which one is best for you.
The RTX 4090 packs a lot more CUDA Cores
Looking at the specs of each GPU, it’s clear that RTX 4090 has a lot more power. While both pieces of hardware use the same Ada Lovelace architecture, the RTX 4090 has 16,384 CUDA Cores, while the RTX 4080 uses either 7,680/9,728 CUDA Cores depending on the SKU you opt for.
Having more CUDA Cores means that the hardware can process more data parallelly. They are essentially the GPU equivalent of CPU cores, with more of them meaning that more information can be processed at once. This is ideal for anyone who wants to play demanding games, especially with features like ray tracing enabled.
While we can’t confirm the performance of these GPUs yet, it looks like the RTX 4090 will have a much smoother, faster and more efficient performance than the RTX 4080, making it the better choice in terms of raw performance.
The RTX 4080 comes in two variations
Nvidia only released two GPUs during the event, although it’s important to note that the RTX 4080 comes in two variations, unlike the RTX 4090, which only comes in one flavour.
The first configuration of the RTX 4080 comes with 12GB of memory, alongside 7,680 CUDA Cores, 639 Tensor-TFLOPs and 92 RT-TFLOPs. The second configuration, which is more expensive, has 16GB memory and comes with 9,728 CUDA Cores, 780 Tensor-TFLOPs and 113 RT-TFLOPs, making it the more powerful option overall.
This means that anyone who wants to look around further than the flagship model can consider both the RTX 4080 GPUs, depending on how much memory and performance power you need.
Both RTX 4080 variations have less memory
Despite the fact that the RTX 4080 comes in two variations, both pack less memory than the RTX 4090. The RTX 4080 comes in either 12GB or 16GB, with the latter costing more.
However, the RTX 4090 comes with 24GB, making it the superior option if you’re worried about memory. It is important to note that all models use GDDR6X memory.
It has improved data transfer rates and is around 43% faster when dealing with data than GDDR6, meaning that despite the lack of memory on the RTX 4080, it should still be able to process data quickly during intensive graphical workloads, like gaming and 3D rendering, although it is unlikely to be as fast as the RTX 4090.
The flagship is more expensive
Unsurprisingly, the RTX 4090 is more expensive than its little brother, the RTX 4080. The RTX 4080 comes in two configurations, with the most powerful flavour starting at $1199, while the other starts at $899.
Meanwhile, the RTX 4090 has only one variation but it costs a lot more than its companions, with a starting price of $1599. While we will run through more specific specs for both GPUs, it’s important to note that you will be spending more money if you’re looking to purchase the flagship model.
The RTX 4090 has faster ray tracing rates
Nvidia introduced 3rd-generation ray tracing on its Lovelace line, meaning that both GPUs will have improved ray tracing capabilities when compared to its predecessors.
However, thanks to the improved specs on the RTX 4090, it will have better performance when dealing with ray tracing, allowing gamers to play supported games with improved frame rates and graphical performance.
The high-end RTX 4080 uses 113 RT-TFLOPS, while the RTX 4090 features 191 RT-TFLOPS. While we won’t be deep diving into what these specs mean specifically, they do allow for more immersive ray tracing. It’s important to note that both GPUs will be capable of ray tracing, although the RTX 4090 will likely have better and faster performance.