2 years ago
34
In a abbreviated station published connected NVIDIA’s website today, the institution has announced that it is “unlaunching” their planned GeForce RTX 4080 12GB card. The lowest-end of the initially denote RTX 40 bid cards, the RTX 4080 12GB had attracted important disapproval since it’s announcement for bifurcating the 4080 tier betwixt 2 cards that didn’t adjacent stock a communal GPU. Seemingly bowing to the unit of those complaints, NVIDIA has removed the paper from their RTX 40 bid lineup, arsenic good arsenic cancelling its November launch.
NVIDIA’s little message reads arsenic follows:
The RTX 4080 12GB is simply a fantastic graphics card, but it’s not named right. Having 2 GPUs with the 4080 designation is confusing.
So, we’re pressing the “unlaunch” fastener connected the 4080 12GB. The RTX 4080 16GB is astonishing and connected way to delight gamers everyplace connected November 16th.
If the lines astir the artifact and enthusiasm for the 4090 is immoderate indication, the reception for the 4080 volition beryllium awesome.
NVIDIA is not providing immoderate further details astir their aboriginal plans for the AD104-based video paper astatine this time. However fixed the circumstances, it’s a tenable presumption close present that NVIDIA present intends to motorboat it astatine a aboriginal time, with a antithetic portion number.
| NVIDIA GeForce Specification Comparison | |||||
| RTX 4090 | RTX 4080 16GB | RTX 4080 12GB (Cancelled) |
|||
| CUDA Cores | 16384 | 9728 | 7680 | ||
| ROPs | 176 | 112 | 80 | ||
| Boost Clock | 2520MHz | 2505MHz | 2610MHz | ||
| Memory Clock | 21Gbps GDDR6X | 22.4Gbps GDDR6X | 21Gbps GDDR6X | ||
| Memory Bus Width | 384-bit | 256-bit | 192-bit | ||
| VRAM | 24GB | 16GB | 12GB | ||
| Single Precision Perf. | 82.6 TFLOPS | 48.7 TFLOPS | 40.1 TFLOPS | ||
| Tensor Perf. (FP16) | 330 TFLOPS | 195 TFLOPS | 160 TFLOPS | ||
| Tensor Perf. (FP8) | 660 TFLOPS | 390 TFLOPS | 321 TFLOPS | ||
| TDP | 450W | 320W | 285W | ||
| L2 Cache | 72MB | 64MB | 48MB | ||
| GPU | AD102 | AD103 | AD104 | ||
| Transistor Count | 76.3B | 45.9B | 35.8B | ||
| Architecture | Ada Lovelace | Ada Lovelace | Ada Lovelace | ||
| Manufacturing Process | TSMC 4N | TSMC 4N | TSMC 4N | ||
| Launch Date | 10/12/2022 | 11/16/2022 | Never | ||
| Launch Price | MSRP: $1599 | MSRP: $1199 | Was: $899 | ||
Taking a look astatine the specifications of the cards, it’s casual to spot wherefore NVIDIA’s halfway basal of enthusiast gamers were not amused. While some RTX 4080 parts shared a communal architecture, they did not stock a communal GPU. Or, for that matter, communal performance.
The RTX 4080 12GB, arsenic it was, would person been based connected the smaller AD104 GPU, alternatively than the AD103 GPU utilized for the 16GB model. In practice, this would person caused the 12GB exemplary to present lone astir 82% of the former’s shader/tensor throughput, and conscionable 70% of the representation bandwidth. A sizable show spread that NVIDIA’s ain figures up of the motorboat person each but confirmed.
NVIDIA, for its part, is nary alien to overloading a merchandise enactment successful this fashion, with similarly-named parts delivering unequal show and the quality denoted solely by their VRAM capacity. This was a signifier that started with the GTX 1060 series, and continued with the RTX 3080 series. However, the show spread betwixt the RTX 4080 parts was acold larger than thing NVIDIA has antecedently done, bringing a bully woody much attraction to the problems that travel from having specified disparate parts sharing a communal merchandise name.
Of adjacent disapproval has been NVIDIA’s determination to merchantability an AD104 portion arsenic an RTX 4080 paper to statesman with. Traditionally successful NVIDIA’s merchandise stack, the adjacent paper beneath the xx80 paper is immoderate signifier of xx70 card. And portion video paper names and GPU identifiers are fundamentally arbitrary, NVIDIA’s aboriginal show figures painted a representation of a paper that would person performed a batch similar the benignant of paper astir radical would expect from the RTX 4070 – delivering show upwards of 20% (or more) down the amended RTX 4080, and on-par with the last-generation flagship, the RTX 3090 Ti. In different words, determination has been a large woody of suspicion wrong the enthusiast assemblage that NVIDIA was attempting to merchantability what different would person been the RTX 4070 arsenic an RTX 4080, portion carrying a higher terms to match.
In immoderate case, those plans are present officially scuttled. Whatever NVIDIA has planned for their AD104-based RTX 40 bid paper is thing lone the institution knows astatine this time. Meanwhile travel November 16th erstwhile the RTX 4080 bid launches, the 16GB AD103-based cards volition beryllium the lone offerings available, with prices starting astatine $1199.
