The Nvidia GeForce RTX 4080 is the follow-up to final month’s RTX 4090 launch, now one of many greatest graphics playing cards and the highest cart in our GPU benchmarks hierarchy. In fact, a little bit of the shine has come off because of the melting 16-pin connectors. The excellent news: RTX 4080 makes use of much less energy, which ought to imply it is also much less more likely to funnel sufficient energy to soften the plastic connector… possibly. The dangerous information: At $1,199 it is nonetheless priced out of attain for many players and represents a giant soar in generational pricing, inheriting the 3080 Ti launch value that we additionally felt was too excessive.
We already know most of what to anticipate from Nvidia’s Ada Lovelace structure, so the one actual query now’s how efficiency scales right down to fewer GPU shaders, much less reminiscence, much less cache, a narrower reminiscence interface, and so forth. Let’s rapidly take a look at the specs for just a few of the highest Nvidia playing cards.
| Graphics Card | RTX 4080 | RTX 4090 | RTX 3090 Ti | RTX 3080 Ti | RTX 3080 | RX 7900 XTX | RX 7900 XT |
|---|---|---|---|---|---|---|---|
| Structure | AD103 | AD102 | GA102 | GA102 | GA102 | Navi 31 | Navi 31 |
| Course of Know-how | TSMC 4N | TSMC 4N | Samsung 8N | Samsung 8N | Samsung 8N | TSMC N5 + N6 | TSMC N5 + N6 |
| Transistors (Billion) | 45.9 | 76.3 | 28.3 | 28.3 | 28.3 | 45.6 + 6x 2.05 | 45.6 + 5x 2.05 |
| Die dimension (mm^2) | 378.6 | 608.4 | 628.4 | 628.4 | 628.4 | 300 + 222 | 300 + 185 |
| SMs | 76 | 128 | 84 | 80 | 68 | 96 | 84 |
| GPU Shaders | 9728 | 16384 | 10752 | 10240 | 8704 | 12288 | 10752 |
| Tensor Cores | 304 | 512 | 336 | 320 | 272 | N/A | N/A |
| Ray Tracing “Cores” | 76 | 128 | 84 | 80 | 68 | 96 | 84 |
| Enhance Clock (MHz) | 2505 | 2520 | 1860 | 1665 | 1710 | 2500 | 2400 |
| VRAM Velocity (Gbps) | 22.4 | 21 | 21 | 19 | 19 | 20 | 20 |
| VRAM (GB) | 16 | 24 | 24 | 12 | 10 | 24 | 20 |
| VRAM Bus Width | 256 | 384 | 384 | 384 | 320 | 384 | 320 |
| L2 Cache | 64 | 72 | 6 | 6 | 5 | 96 | 80 |
| ROPs | 112 | 176 | 112 | 112 | 96 | 192 | 192 |
| TMUs | 304 | 512 | 336 | 320 | 272 | 384 | 336 |
| TFLOPS FP32 | 48.7 | 82.6 | 40 | 34.1 | 29.8 | 61.4 | 51.6 |
| TFLOPS FP16 (FP8/INT8) | 390 (780) | 661 (1321) | 160 (320) | 136 (273) | 119 (238) | 123 (246) | 103 (206) |
| Bandwidth (GBps) | 717 | 1008 | 1008 | 912 | 760 | 960 | 800 |
| TBP (watts) | 320 | 450 | 450 | 350 | 320 | 355 | 300 |
| Launch Date | Nov 2022 | Oct 2022 | Mar 2022 | Jun 2021 | Sep 2020 | Dec 2022 | Dec 2022 |
| Launch Value | $1,199 | $1,599 | $1,999 | $1,199 | $699 | $999 | $899 |
There is a comparatively massive hole between the RTX 4080 and the bigger RTX 4090. You get most of an AD103 GPU — 76 of the potential 80 Streaming Multiprocessors (SMs) — however that is nonetheless 40% fewer GPU shaders and different purposeful models than the RTX 4090. Clock speeds are related, you get 33% fewer reminiscence channels, VRAM, and bandwidth, and the rated TBP drops by 29%. On paper, the RTX 4090 could possibly be as much as 70% sooner based mostly on the theoretical compute efficiency, and that is a priority.
$1,199 is hardly inexpensive, so if seems like anybody even wanting on the RTX 4080 ought to in all probability simply save up the extra $400 for the RTX 4090 and go for broke — or melted. However then the RTX 4090 has been bought out at wherever beneath $2,100 since launch, which implies it might really be a $900 upsell, and that is way more important.
The pricing turns into much more of a priority after we consider AMD’s Radeon RX 7900 XTX/XT playing cards coming subsequent month. We now have all of the pertinent particulars for the primary playing cards utilizing AMD’s RDNA 3 GPU structure, they usually actually look promising. Costs are nonetheless excessive, however the specs comparisons recommend AMD may be capable of beat the RTX 4080 whereas costing not less than $200–$300 much less. Which implies, except you completely refuse to think about buying an AMD graphics card, it’s best to not less than wait till subsequent month to see what the crimson crew has to supply.
Picture 1 of 2
Nvidia does have some extras that AMD is unlikely to match within the close to time period, nevertheless. For instance, the Deep studying and AI horsepower in RTX 4080 far surpasses what AMD intends to supply. If we have got the figures proper, AMD’s FP16 and INT8 throughput shall be lower than a 3rd of the RTX 4080.
Nvidia additionally provides DLSS 3 courtesy of the improved Optical Movement Accelerator (OFA), and there are already ten video games that assist the know-how: Brilliant Reminiscence: Infinite, Destroy All People! 2 – Reprobed, F.I.S.T.: Solid in Shadow Torch, F1 22, Justice, Loopmancer, Marvel’s Spider-Man Remastered, Microsoft Flight Simulator , A Plague Story: Requiem, and Tremendous Folks. That is about half as many DLSS 3 video games in lower than a month as there are video games with AMD’s FSR2 know-how. In fact, you want an RTX 40-series GPU for DLSS 3, whereas FSR2 works with just about every little thing.
Nvidia GPUs additionally are typically closely favored by skilled customers, or not less than their employers. Whereas true workstations will doubtless go for the RTX 6000 48GB card versus a GeForce RTX 40-series, there is definitely potential in selecting up a number of RTX 4080 playing cards for AI and deep studying use. Content material creators may discover one thing to love, although once more, when you’re prepared to pay for a 4080, it will not be an enormous step up in pricing to nab a 4090 as an alternative.
One other piece of fine information (relying on which aspect of the aisle you fall, we suppose) is that GPU mining stays unprofitable. Players will not be capable of offset the worth of a brand new graphics card by cryptocurrency mining, however not less than there needs to be extra GPUs accessible for players. Now let’s examine precisely what Nvidia has to supply with its new RTX 4080.