The memory specifications of the upcoming RTX 3090 and RTX 3080 Ampere cards have just been confirmed by Videocardz. According to Videocardz’s internal sources at Add-in-Board (AIB) Partners, it has been confirmed that the NVIDIA GeForce RTX 3090 will feature 24 GB of GDDR6X memory, while the GeForce RTX 3080 will initially launch with 10 GB of GDDR6X memory. There’s a huge difference between the RTX 3090 & the RTX 3080 when it comes to the VRAM capacity, but only these specs have been confirmed for now.
There is no word on any Ti Model, but the RTX 3090 appears to be the flagship GPU in the Ampere lineup, directly replacing TITAN. The RTX 3090 will feature the GA102-300-A1 GPU. The GA102-300-A1 GPU sports 5248 CUDA cores or 82 SMs, which is a 20% increase in cores over the RTX 2080 Ti. We don’t have any details on the TMU (texture mapping unit), ROP count (raster operations pipeline) and the clock speeds yet. Assuming the RTX 3090 comes with 24 GB of VRAM then we are looking at a 384-bit wide memory bus, and 1 TB/s of bandwidth. The memory speeds for both the RTX 3090 and 3080 cards are expected to be around 19 Gbps.
But it is important to note that Gaming and HPC GPU variants have different configurations, so we can’t take the GA102-300-A1 GPU die as a reference if the RTX 3090 is going to target both the consumer and the HPC/enterprise market. The GeForce RTX 3080 on the other hand has been confirmed to get 10 GB GDDR6X memory, which appears to less that the RTX 3090, but this could also mean that we might get an RTX 3080 SUPER variant GPU with 16 GB of VRAM in the future. These super cards might also come in 20 GB flavors but for now only 10 GB of memory has been confirmed.
There are chances that Nvidia will launch the RTX SUPER series next year though. The RTX 3080 will feature the GA102-200-KD-A1 GPU die, which is a cut-down SKU having the same 4352 CUDA cores as the RTX 2080 Ti with a total of 68 SMs. Assuming the memory is running at 19 Gbps across a 320-bit wide bus interface, we can expect a bandwidth of up to 760 GB/s.
Nvidia is using the high-end GA102 GPU for the RTX 3080 as well, which appears to be an upgrade over the previous TU104 core featured on the RTX 2080. This could also mean that the new Ampere card will have higher wattage requirements and thermals, and it will fall under the high-end enthusiast segment. The GeForce RTX 3070 is also expected to be announced along with these two high-end cards, but we don’t have any specs on this SKU yet.
NVIDIA’s flagship RTX 3090 graphics card was also pictured recently, and is a massive 3-slot GPU. This model appears to have the same irregular shaped PCB which was leaked before. My guess is that this is the reference Founders Edition graphics card sample.
3 slots. 2000$. pic.twitter.com/ZgKsPUQ1U4
— James (@GarnetSunset) August 21, 2020
NVIDIA’s upcoming RTX 3000 series Ampere cards are rumored to be built on the Samsung’s 8nm process node. Nvidia has also recently ceased the production of the RTX 20 series of GPUs to make room for the Ampere lineup. Nvidia is hosting a Geforce Special Event on September 1st, and we expect the company to announce the next-gen Ampere Gaming GPUs.
Stay tuned for more!