With a lot anticipation and quite a lot of leaks, NVIDIA this morning is asserting the subsequent technology of video playing cards, the GeForce RTX 30 collection. Based upon the gaming and graphics variant of NVIDIA’s Ampere structure and constructed on an optimized model of Samsung’s 8nm course of, NVIDIA is touting the brand new playing cards as delivering a few of their biggest beneficial properties ever in gaming efficiency. All the whereas, the most recent technology of GeForce may even be coming with some new options to additional set the playing cards other than and forward of NVIDIA’s Turing-based RTX 20 collection.
Out of the gate, NVIDIA is asserting the primary three playing cards to make up the brand new RTX 30 collection: the RTX 3090, RTX 3080, and RTX 3070. These playing cards are all launching inside the subsequent month and a half – albeit at barely separate instances – with the RTX 3090 and RTX 3080 main the cost. The two playing cards, in flip, will function the successors to NVIDIA’s GeForce RTX 2080 Ti and RTX 2080/2080S respectively, hitting new highs in graphics efficiency, albeit whereas additionally hitting new highs in costs within the case of the RTX 3090.
The first card out the door would be the GeForce RTX 3080. With NVIDIA touting upwards of 2x the efficiency of the RTX 2080, this card will go on sale on September 17th for $700. That can be adopted up per week later by the much more highly effective GeFoce RTX 3090, which hits the cabinets September 24th for $1500. Finally, the RTX 3070, which is being positioned as extra of a conventional candy spot card, will arrive subsequent month at $499.
NVIDIA GeForce Specification Comparison | ||||||
RTX 3090 | RTX 3080 | RTX 3070 | RTX 2080 Ti | |||
CUDA Cores | 10496 | 8704 | 5888 | 4352 | ||
Boost Clock | 1.7GHz | 1.71GHz | 1.73GHz | 1545MHz | ||
Memory Clock | 19.5Gbps GDDR6X | 19Gbps GDDR6X | 16Gbps GDDR6 | 14Gbps GDDR6 | ||
Memory Bus Width | 384-bit | 320-bit | 256-bit | 352-bit | ||
VRAM | 24GB | 10GB | 8GB | 11GB | ||
Single Precision Perf. | 35.7 TFLOPs | 29.eight TFLOPs | 20.four TFLOPs | 13.four TFLOPs | ||
Tensor Perf. (FP16) | 285 TFLOPs | 238 TFLOPs | 163 TFLOPs | 114 TFLOPs | ||
Ray Perf. | 69 TFLOPs | 58 TFLOPs | 40 TFLOPs | ? | ||
TDP | 350W | 320W | 220W | 250W | ||
GPU | GA102? | GA102? | GA104? | TU102 | ||
Transistor Count | 28B | 28B | ? | 18.6B | ||
Architecture | Ampere | Ampere | Ampere | Turing | ||
Manufacturing Process | Samsung 8nm | Samsung 8nm | Samsung 8nm | TSMC 12nm “FFN” | ||
Launch Date | 09/24/2020 | 09/17/2020 | 10/2020 | 09/20/2018 | ||
Launch Price | MSRP: $1499 | MSRP: $699 | MSRP: $499 | MSRP: $999 Founders $1199 |
Ampere for Gaming: GA102
As is historically the case for NVIDIA, this morning’s public presentation was not an architectural deep dive. Though the purely digital presentation was definitely a change of tempo for an organization who treats each video card launch like a celebration, NVIDIA caught to their profitable launch playbook. That means quite a lot of demonstrations, testimonials, and promotional movies, together with some high-level overviews of a number of of the applied sciences and engineering design selections that went into making their newest technology of GPUs. The internet result’s that we now have an honest thought of what’s in retailer for the RTX 30 collection, however we’ll have to attend for NVIDIA to supply some deep dive technical briefings to fill within the blanks and get to the guts of issues in true AnandTech model.
At a excessive stage, Ampere and the GA102 GPU being utilized in these top-tier playing cards brings a number of main {hardware} developments to NVIDIA’s lineup. The largest of which is the ever-shrinking measurement of transistors, because of a custom-made model of Samsung’s 8nm course of. We solely have restricted details about this course of – principally as a result of it hasn’t been used too many locations – however at a excessive stage it is Samsung’s densest…