Alleged specifications of NVIDIA's next-generation GeForce RTX 3090, RTX 3080 & Titan RTX with Ampere Gaming GPUs have been detailed. The NVIDIA GeForce RTX 30 series will be featuring a brand new graphics architecture and other key driving features for faster gaming performance in next-generation AAA titles.
NVIDIA GeForce RTX 3090, RTX 3080 & Titan RTX With Ampere Gaming Graphics Cards Allegedly Feature GA102 GPUs
There are several rumors going on around the GeForce RTX 30 series since the coolers leaked out last month. But almost everywhere we're hearing one thing consistently which is that the GeForce RTX 3080 and GeForce RTX 3090 will be the first cards featuring the Ampere Gaming GPUs. Latest version of specifications come from KatCorgi and Kopite7kimi (via Videocardz) who have not only shared the same specs in the current month but Kopite7kimi was accurate with his previous leaks for the A100 GPU.
Based on that, let's see what latest details do these two leakers have to share with us but most importantly, since all of this is unconfirmed at the moment, it will be wise to treat this rumor with a grain of salt until we recieve more information.
NVIDIA GeForce RTX 3090 - The Fastest Ampere Gaming Graphics Card
The NVIDIA GeForce RTX 3090 is an interesting graphics card based on its naming scheme. I already talked around the nomenclature of this card in my previous post mentioning why NVIDIA would be choosing this particular naming scheme that was previously reserved for dual-chip graphics cards. It is highly likely that the RTX 3090 could be an internal codename for the RTX 3080 Ti but if we dive to much into those details, we will just be moving away from today's topic.
So let's call it the GeForce RTX 3090 for now. According to the details, the GeForce RTX 3090 will feature the GA102-300-A1 GPU. This is one of the three SKUs that we will be talking about in this particular post. The GA102-300-A1 GPU is said to be comprised of 5248 CUDA cores or 82 SMs. In total, that's a 20% increase in cores over the GeForce RTX 2080 Ti. No details such as clock speeds, TMU/ROP counts is provided and we can't take the Ampere A100 die as a reference since gaming and HPC parts share a different configuration hierarchy.
Moving on to the memory side, things start to look interesting as it is stated that the card will get up to 12 GB memory with speeds of up to 21 Gbps. Since that's a 384-bit bus we are looking at, it will be able to offer up to 1 TB/s bandwidth. This overclocked memory design is being referred to as GDDR6X and there's no such information available yet by any graphics DRAM vendor unlike the previous GDDR6 DRAM which was officially reported months before its integration on the Turing graphics cards.
NVIDIA GeForce RTX 3080 - GA102 For High-End Gaming
Moving on to the GeForce RTX 3080, the rumor reports that the card will be featuring the GA102-200-Kx-A1 SKU. This cut down SKU will feature the same 4352 CUDA cores as the RTX 2080 Ti that will be arranged in a total of 68 SMs. The card is reportedly going to feature 10 GB of memory that is also going to be GDDR6X, running at 19 Gbps across a 320-bit bus interface with a bandwidth of 760 GB/s.
The switch to GA102 from TU104 on the RTX 2080 is definitely an interesting one. A high-end SKU would also result in higher wattage and thermals but I believe this could be done to raise the bar up on the sub $500 segment which should comprise of GA104 based RTX 3070 and RTX 3060. If that's the case, than we can expect a good performance jump in that segment which we recently talked about in one of our article.
NVIDIA 2nd Gen Titan RTX - Based on Unlocked Ampere GA102 GPU
Finally, we have reports on a 2nd Generation Titan graphics cards. The card is stated to be a 2nd generation Titan RTX card but like the Titan X and Titan X(p) before it, the new card would make use of the Ampere GA102-400-A1 GPU. The GPU will be composed of 84 SMs for a total of 5376 CUDA cores.
The card is said to feature 24 GB of GDDR6X memory across a 384-bit bus interface and a memory bandwidth of 816 GB/s which although is lower than the RTX 3090 due to a slower pin speed of 17 Gbps but will be compensated by the enhanced memory compression architecture plus you will be getting twice the memory buffer as the RTX 3090.
NVIDIA GeForce RTX 30 Series 'Ampere Gaming GPU' Rumored Specifications:
| Graphics Card Name | GPU | Cores (SMs) | Memory | Memory Bus | Bandwidth |
|---|---|---|---|---|---|
| NVIDIA Titan RTX (2nd Gen) | Ampere GA102-400 | 5376 (84) | 24 GB GDDR6X | 384-bit | ~816 GB/s |
| NVIDIA Titan RTX (1st Gen) | Turing TU102-400 | 4608 (72) | 24 GB GDDR6 | 384-bit | 672 GB/s |
| NVIDIA GeForce RTX 3080 Ti / NVIDIA GeForce RTX 3090 | Ampere GA102-300 | 5248 (82) | 12 GB GDDR6X | 384-bit | ~1.00 TB/s |
| NVIDIA GeForce RTX 2080 Ti | Turing TU102-300 | 4352 (68) | 11 GB GDDR6 | 352-bit | 616 GB/s |
| NVIDIA GeForce RTX 3080 | Ampere GA102-200 | 4352 (68) | 10 GB GDDR6X | 320-bit | ~760 GB/s |
| NVIDIA GeForce RTX 2080 | Turing TU104-400 | 2944 (46) | 8 GB GDDR6 | 256-bit | 448 GB/s |
| NVIDIA GeForce RTX 3070 Ti | Ampere GA104-400 | 3072 (48) | 8 GB GDDR6X | 256-bit | ~512 GB/s |
| NVIDIA GeForce RTX 2070 SUPER | Turing TU104-410 | 2560 (40) | 8 GB GDDR6 | 256-bit | 448 GB/s |
| NVIDIA GeForce RTX 3070 | Ampere GA104-300 | 2944 (46) | 8 GB GDDR6 | 256-bit | ~512 GB/s |
| NVIDIA GeForce RTX 2070 | Turing TU106-400 | 2304 (36) | 8 GB GDDR6 | 256-bit | 448 GB/s |
| NVIDIA GeForce RTX 3060 Ti | Ampere GA106-400 | TBD | TBD | TBD | TBD |
| NVIDIA GeForce RTX 2060 SUPER | Turing TU106-410 | 2176 (34) | 8 GB GDDR6 | 256-bit | 448 GB/s |
| NVIDIA GeForce RTX 3060 | Ampere GA106-300 | TBD | TBD | TBD | TBD |
| NVIDIA GeForce RTX 2060 | Turing TU106-200 | 1920 (30) | 6 GB GDDR6 | 192-bit | 336 GB/s |
According to Jensen, all next-generation products will be featuring the same NVIDIA Ampere GPU architecture. The first outing of Ampere will be for the cloud and research segment in the form of the Ampere GA100 GPU which is designed solely for the HPC environment. The gaming cards will definitely feature Ampere GPUs but will come later.
Ampere will eventually replace Nvidia’s Turing and Volta chips with a single platform that streamlines Nvidia’s GPU lineup, Huang said in a pre-briefing with media members Wednesday. While consumers largely know Nvidia for its videogame hardware, the first launches with Ampere are aimed at AI needs in the cloud and for research.
“Unquestionably, it’s the first time that we’ve unified the acceleration workload of the entire data center into one single platform,” Huang said. - Market Watch
NVIDIA's GeForce RTX 30 series is expected to launch in the coming months with September being touted as a potential launch window. On the other hand, AMD is also planning a big Radeon revival with its upcoming Big Navi graphics cards which are also expected around late Q3 2020.
