If you have been waiting for Nvidia’s upcoming 3000 series GPUs, then you know that the time draws neigh onto the official release of these powerful new GPUs – the RTX 3070, 3080 and 3090. But since no company is immune to leaks, and especially not Nvidia, we have been receiving a slow stream of information regarding the three first GPUs to be released toward the end of September: The current flagship, GeForce RTX Ampere 3090, and the position two GeForce RTX Ampere 3080. Now that we have the actual specifications, let’s compare and analyze the two and see which model best fits your needs; which is a matter of whether the RTX 3090 is worth its increased price tag, or if the RTX 3080 GPU displays enough value to fulfill your gaming needs.
RTX 3090 vs 3080 – Specification Comparison
|RTX 3090||RTX 3080|
|Boost clock||1700 MHz||1710 MHz|
|Total VRAM||24GB GDDR6X||10B GDDR6X|
|Single-precision throughput||36 TFLOPs||30 TFLOPs|
|Node||Samsung 8NM||Samsung 8NM|
|Connectors||1x HDMI2.1, 3x DP1.4a||1x HDMI2.1, 3x DP1.4a|
|Launch MSRP USD||$1499||$699|
RTX 3090 vs 3080: Commonalities
But before we get into the comparison of the RTX 3080 and RTX 3090, let’s take a look at what they have in common. Firstly, both models (as well as, most likely, all 3000 series GPUs) will have second-generation Ray Tracing Cores and third-generation Tensor cores. Next, both models are set to use variants Nvidia’s GA102 GPU, which makes use of the Ampere architecture and an 8nm Samsung fabrication process. This GPU has a die size of 627 mm², 28 billion transistors, and will support DirectX 12 Ultimate. Another common feature of these two models is their use of the GDDR6X VRAM. Micron’s GDDR6X offers both an increased memory bandwidth of 912 to 1,008 GB/s (over the 672 to 768 GB/s memory bandwidth of the previous generation VRAM) as well as a slight power improvement in comparison to the GDDR6 (7.25 pJp vs 7.5 pJp). Both GPUs will also feature PCI Express Gen 4, as well as Nvidia’s DLSS, GeForce Experience, GPU Boost, and G-Sync software. Finally, the RTX 3080 and the RTX 3090 will feature three DisplayPort 1.4a ports, as well as a HDMI 2.1 ports – though the number of ports will ultimately be aftermarket dependent.
RTX 3090 vs 3080: Memory
Where the RTX 3080 got shortchanged the most, in comparison to its RTX 3090 counterpart, is its VRAM capacity. The GeForce RTX 3090 will come equipped with 24 GB of GDDR6X VRAM, versus the only 10 GBs of the RTX 3080: a substantial difference. The 10 GB VRAM of the RTX 3080 also puts it behind the current generation RTX 2080 Ti which has 11 GB of VRAM; though the RTX 3080’s use of GDDR6X memory means that its VRAM will have a faster bandwidth and better power management. But when compared to the RTX 3090, the RTX 3080 GPU will trail, not only in the size of its VRAM, but also its quality. The bus width of the RTX 3080’s memory is 320-bit, whereas the RTX 3090’s memory has a 384-bit bus width. Also, the max bandwidth and memory for the latter GPU is 936 GB/s and 9,750 MHz, which is superior to the 760 GB/s bandwidth and 9,500 MHz of the RTX 3080. When compared to the previous flagship 2080 Ti, as well as the 2080 Super, both these GPUs offer far more VRAM storage that is of higher quality than their predecessors. Of course, if the rumors are true about a potential RTX 3080 Ti with a VRAM of 12 GB, the difference with the previous generation becomes a bit less impressive, at least in terms of RAM.
RTX 3090 vs 3080: Graphical Processing Power
As previously mentioned, both the RTX 3080 and RTX 3090 are set to be released with a GA102, 8nm, GPU – though the 3090 will use a GA102-300 on a PG132 SKU 30 board, whereas the 3080 will use the GA102-200 on a PG132 SKU 10 board.
The number of CUDA Cores of the 3090 are 10,496, with 328 TMUs and 96 ROPs – an 140% increase in comparison to the 4,352 CUDA Cores of the RTX 2080 Ti.
The number of CUDA Cores on the 3080 will be 8,704 and the TMUs and ROPs may be 272 and 88 respectively. Again, an 100% increase in CUDA cores when compared to the current-gen RTX 2080 Ti. In terms of TFLOPS, the RTX 3090 is set to release with 36 Shader-FLOPS, 69 RT-TFLOPS, and 285 Tensor-TFLOPS, whereas the RTX 3080 will have 30 Shader-TFLOPS, 58 Tensor-TFLOPS, and 238 Shader-TFLOPS.
Finally, the base and boost clock speeds of the RTX 3090’s reference card are 1,4 MHz and 1,7 MHz respectively, whereas the RTX 3080’s reference card displays a slightly higher base clock of 1.44 MHz and boost clock speed of 1,710 MHz. Of course, the boost clock speeds will vary, depending on aftermarket manufacturer.
RTX 3090 vs 3080: Power Consumption
Both the RTX 3080 and RTX 3090 will be quite power-hungry, with 320W and 350W TDPs respectively, so a 750W PSU will be the minimum suggested power supply for both these Nvidia GeForce Ampere 3000 Series GPUs, and ideally, an 850W PSU will be most suitable for the 3090.
RTX 3090 vs 3080: Price & Overall Value Comparison
Though we do see that the RTX 3090 has substantially superior specifications when compared to the RTX 3080… we still have to wonder: Is the RTX 3090 worth its price?
According to the official announcements by Nvidia, the RTX 3090 is set to release with a $1,399 price tag, whereas the RTX 3080 is set to release with a $799 price tag. A difference of $600. So now the question becomes: Is having double the amount of memory, and about a 20% increase in CUDA Cores really worth $600?
Well, the answer really depends on what kind of gaming setup you are looking to make, and what type of gaming you prefer. For multiplayer competitive games, both these GPUs are overkill, so the question is whether you are content in playing triple-A games at 144 FPS, and if so, on how high of a resolution. With the info we have now, it is looking like the Nvidia GeForce Ampere RTX 3080 will perform can easily reach the 144 Hz mark on a 1440p monitor, a feat that the generation GeForce RTX 2080 Ti could not accomplish – despite having a $400 higher price tag. Still, playing triple-A games at 144 Hz, on maximum settings, and resolutions above 1440p may not always be possible. The RTX 3080 will definitely be able to handle any game (current or next gen) at high resolutions, and at well above 60 FPS, but at 144 Hz the graphical processing power demanded is extremely high, proven by the performance of the RTX 2080 Ti. A great example is Call of Duty: Warzone. Benchmarks carried out by ComputerBase with an Intel i9-9900K CPU and top of the line RAM show us that the RTX 2080 Ti FE was unable to reach a framerate above 100 FPS at 1440p. Whether or not the RTX 3080 will be capable of pulling off over 120 FPS for 4K resolution with its new 8nm Ampere architecture and GDDR6X VRAM is yet to be seen; but until we await the benchmarks, it must be noted than next-generation games will become more and more hardware demanding, so we can’t be sure if it will be able to sustain such framerates and at such high resolutions in the future.
On the other hand, the GeForce RTX Ampere 3090 seems poised to be the ultimate GPU for gaming… or for any other type of application, really. Of course, how well it will perform in comparison to the RTX 2080 Ti or the RTX 3080 is something we’ll have to wait in order to see, though on paper it does show promise as being the 144 Hz, 4K, GPU champion that we’ve been waiting for. Also, Nvidia has claimed that the RTX 3090 will able to play games at 8K resolutions at above 60 FPS, but that is also something that we need to see in benchmarks. If it is able to do so, though, then it will definitely be worth saving for, as the luxurious RTX 3090 is going to cost a pretty penny.