The second main keynote of the day for Computex comes from NVIDIA, who has taken to the digital showfloor to debate new merchandise for each avid gamers and the enterprise market. On the gaming aspect of issues, the corporate is formally saying its GeForce RTX 3080 Ti and RTX 3070 Ti playing cards. The refreshed GeForce playing cards will land in early June, additional increasing the high-end of NVIDIA’s video card lineup.
The newest members of the GeForce household are arriving at a little bit of an odd time. NVIDIA’s present lineup of high-end playing cards launched not even 9 months in the past, in order that they’re not by typical metrics due for a mid-generation kicker, resembling what we noticed with 2019’s RTX 20 Super collection. Nor is NVIDIA having any downside promoting the playing cards they’ve – due to excessive demand they’re promoting all the things they make, and are making a pleasant revenue margin whereas doing so. None the much less, NVIDIA is utilizing Computex to additional broaden the GeForce household lineup with two extra playing cards, the RTX 3080 Ti and RTX 3070 Ti.
Both playing cards are based mostly on current GPUs and applied sciences, so with no intention of downplaying each playing cards, there’s nothing novel right here. Instead, these playing cards give NVIDIA and its board companions a couple of extra choices to play with. And for the NVIDIA ecosystem, it permits NVIDIA to have some further playing cards to place towards AMD’s Radeon RX 6000 lineup.
NVIDIA GeForce Specification Comparison | |||||||
RTX 3090 | RTX 3080 Ti | RTX 3080 | RTX 3070 Ti | RTX 3070 | |||
CUDA Cores | 10496 | 10240 | 8704 | 6144 | 5888 | ||
ROPs | 112 | 112 | 96 | 96 | 96 | ||
Boost Clock | 1.7GHz | 1.67GHz | 1.71GHz | 1.77GHz | 1.725GHz | ||
Memory Clock | 19.5Gbps GDDR6X | 19Gbps GDDR6X | 19Gbps GDDR6X | 19Gbps GDDR6X | 14Gbps GDDR6 | ||
Memory Bus Width | 384-bit | 384-bit | 320-bit | 256-bit | 256-bit | ||
VRAM | 24GB | 12GB | 10GB | 8GB | 8GB | ||
Single Precision Perf. | 35.7 TFLOPS | 34.1 TFLOPS | 29.eight TFLOPS | 21.7 TFLOPS | 20.three TFLOPS | ||
Tensor Perf. (FP16) | 143 TFLOPS | 136 TFLOPS | 119 TFLOPS | 87 TFLOPS | 81 TFLOS | ||
Tensor Perf. (FP16-Sparse) | 285 TFLOPS | 273 TFLOPS | 238 TFLOPS | 174 TFLOPS | 163 TFLOPS | ||
TDP | 350W | 350W | 320W | 290W | 220W | ||
GPU | GA102 | GA102 | GA102 | GA104? | GA104 | ||
Transistor Count | 28B | 28B | 28B | 17.4B | 17.4B | ||
Architecture | Ampere | Ampere | Ampere | Ampere | Ampere | ||
Manufacturing Process | Samsung 8nm | Samsung 8nm | Samsung 8nm | Samsung 8nm | Samsung 8nm | ||
Launch Date | 09/24/2020 | 06/03/2021 | 09/17/2020 | 06/10/2021 | 10/29/2020 | ||
Launch Price | MSRP: $1499 | MSRP: $1199 | MSRP: $699 | MSRP: $599 | MSRP: $499 |
GeForce RTX 3080 Ti: The New Gaming Flagship
Starting issues off, now we have the GeForce RTX 3080 Ti. NVIDIA is dubbing this card “the new gaming flagship,” and on the entire is treating it like a standard xx80 Ti card, flagship standing and all. That mentioned, not like the 2080 Ti and 1080 Ti, that is really a barely cut-down model of a greater NVIDIA GeForce card that’s already in the marketplace, the GeForce RTX 3090. So the entire matter is a bit awkward, and it comes with the impression that some a part of NVIDIA want to retroactively make the RTX 3090 a Titan card reasonably than a GeForce card.
But bizarre product positioning apart, what isn’t up for debate is RTX 3080 Ti’s {hardware} credentials. For all sensible functions it is a cheaper RTX 3090 with half of the VRAM, will will make it a really potent video card for 4K gaming. NVIDIA is utilizing the identical GA102 GPU utilized in RTX 3090, disabling a pair extra SMs, dialing down the clockspeeds a hair, and sending it out the door. The ensuing card has 80 SMs enabled (10240 CUDA cores) and is clocked at 1.67GHz. On paper this provides it 96% of the RTX 3090’s compute/shader…