Nvidia made the RTX 4080 official at its GeForce Beyond event, and now that we’ve got some particulars, it’s time to see how the next-generation Ada Lovelace card compares to the previous guard: the RTX 3080. Though we don’t have a variety of concrete particulars but, we nonetheless have sufficient to check the specs, pricing, and anticipated efficiency of the RTX 3080 and RTX 4080.
Pricing and availability
The RTX 3080 launched on September 17, 2020, for an inventory worth of $700. Only a few playing cards truly offered for that worth, although, and the cardboard remained costly for almost two years. Now, you could find fashions out there for round $750, however we anticipate these costs to drop as RTX 40-series playing cards change into extra frequent.
The RTX 4080 launched at Nvidia’s particular GeForce Past broadcast on September 20, 2022. Nvidia hasn’t confirmed a launch date, nevertheless it says the playing cards will arrive someday in November. There are literally two fashions out there, one with 16GB of reminiscence and one other with 12GB. The 16GB mannequin will run $1,200, whereas the 12GB is barely cheaper at $900. Remember that these are the costs set by Nvidia, and board associate playing cards will promote for extra.
There’s a second model of the RTX 3080 as effectively, one with 12GB of reminiscence and a slight enhance to cores. The 12GB RTX 3080 is barely dearer than the bottom mannequin at round $800 to $900, however Nvidia by no means set an official record worth on this card. You’ll usually discover it as simply an RTX 3080 and typically with a barely larger markup.
For present costs, it’s higher to go together with the RTX 3080, however we’ve got to see the place costs land as soon as the RTX 4080 has lastly launched.
Specs
For uncooked specs, essentially the most fascinating notice concerning the two RTX 4080 fashions is that they arrive with a large enhance to clock velocity. Core counts are up a hair, no less than on the 16GB mannequin, however the 800MHz enhance is the large kicker, particularly contemplating the RTX 4080 comes with chipmaker TSMC’s 5nm node.
| RTX 3080 | RTX 4080 16GB | RTX 4080 12GB | |
| Course of | Samsung 8nm | TSMC 5nm | TSMC 5nm |
| Structure | Ada Lovelace | Ampere | Ampere |
| CUDA cores | 8960 / 8704 | 9,728 | 7,680 |
| Reminiscence | 12GB / 10GB GDDR6X | 16GB GDDR6X | 12GB GDDR6X |
| Enhance clock velocity | 1710MHz | 2505MHz | 2610MHz |
| Bus width | 384-bit / 320-bit | 256-bit | 192-bit |
| Energy | 350W /320W | 320W | 285W |
Samsung’s 8nm manufacturing course of did wonders for the Ampere structure within the earlier technology, however manufacturing expertise has moved on. TSMC’s 5nm course of can pack much more energy in the identical space, so though the specs don’t appear like a large uplift on paper, they need to be taken within the context of the brand new node.
One spec we will evaluate is reminiscence bandwidth. Nvidia is sticking with the identical GDDR6X, nevertheless it’s reducing the bus width whereas upping the capability. It’s an fascinating transfer that Nvidia experimented with within the earlier technology with the RTX 3060. It’s unlikely that bandwidth will likely be a giant drawback on these flagship playing cards, nevertheless it’s positively one thing to remember with lower-end RTX 40-series choices.
Nvidia didn’t improve energy calls for, regardless of what rumors claimed. Actually, the 12GB RTX 4080 drops energy calls for to 285W, which is a good minimize. Energy calls for have been trending upwards, so it’s good to see them go in the wrong way this technology.
Anticipated efficiency
Nvidia hasn’t shared many particulars concerning the RTX 4080 but, so we don’t know the way it will carry out. Nonetheless, the corporate says that it’s two to 4 instances sooner than the RTX 3080 Ti, which itself is round 10% to fifteen% sooner than the RTX 3080. You shouldn’t take these numbers as legislation, nonetheless, as a result of the RTX 4080 has some particular tech beneath the hood.
It has two issues specifically: DLSS 3 and Shader Execution Reordering (SER). In contrast to DLSS 2, the brand new model truly predicts new frames. A number of the frames in your recreation will likely be fully generated with out your GPU rendering a single pixel. SER boosts ray tracing efficiency by prioritizing ray tracing operations based mostly on the computational energy out there on the time.
The issue with Nvidia’s efficiency declare is that it’s not clear if SER and DLSS are factored into them. We anticipate a strong uplift in efficiency over the RTX 3080, however with out concrete benchmarks to check, it’s laborious to say a lot at this level.
Go for the deal or efficiency?
Though the RTX 4080 could present a giant improve in efficiency over the RTX 3080, it comes at a giant value. At least, you’re a $200 upcharge and upwards of a $500 improve in worth. And that’s with present RTX 3080 costs, which is able to virtually definitely drop in worth as soon as the RTX 4080 is right here.
All of it comes right down to efficiency, and for now, we’ve got to attend till benchmarks are circulating. In case you’re going for the worth play, although, the RTX 3080 nonetheless appears like a winner.
Editors’ Suggestions
The post Nvidia RTX 4080 vs. RTX 3080: Is Ada Lovelace worth it? | Digital Trends appeared first on NO INDEX.