
In a brief publish printed on NVIDIA’s web site at this time, the corporate has introduced that it’s “unlaunching” their deliberate GeForce RTX 4080 12GB card. The bottom-end of the initially announce RTX 40 sequence playing cards, the RTX 4080 12GB had attracted vital criticism because it’s announcement for bifurcating the 4080 tier between two playing cards that didn’t even share a typical GPU. Seemingly bowing to the strain of these complaints, NVIDIA has eliminated the cardboard from their RTX 40 sequence lineup, in addition to cancelling its November launch.
NVIDIA’s brief message reads as follows:
The RTX 4080 12GB is a implausible graphics card, but it surely’s not named proper. Having two GPUs with the 4080 designation is complicated.So, we’re urgent the “unlaunch” button on the 4080 12GB. The RTX 4080 16GB is superb and on monitor to please avid gamers in every single place on November sixteenth.
If the traces across the block and enthusiasm for the 4090 is any indication, the reception for the 4080 will probably be superior.
NVIDIA shouldn’t be offering any additional particulars about their future plans for the AD104-based video card presently. Nonetheless given the circumstances, it’s an inexpensive assumption proper now that NVIDIA now intends to launch it at a later time, with a unique half quantity.
NVIDIA GeForce Specification Comparability | |||||
RTX 4090 | RTX 4080 16GB | RTX 4080 12GB (Cancelled) |
|||
CUDA Cores | 16384 | 9728 | 7680 | ||
ROPs | 176 | 112 | 80 | ||
Enhance Clock | 2520MHz | 2505MHz | 2610MHz | ||
Reminiscence Clock | 21Gbps GDDR6X | 22.4Gbps GDDR6X | 21Gbps GDDR6X | ||
Reminiscence 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 Depend | 76.3B | 45.9B | 35.8B | ||
Structure | Ada Lovelace | Ada Lovelace | Ada Lovelace | ||
Manufacturing Course of | TSMC 4N | TSMC 4N | TSMC 4N | ||
Launch Date | 10/12/2022 | 11/16/2022 | By no means | ||
Launch Value | MSRP: $1599 | MSRP: $1199 | Was: $899 |
Looking on the specs of the playing cards, it’s straightforward to see why NVIDIA’s core base of fanatic avid gamers weren’t amused. Whereas each RTX 4080 elements shared a typical structure, they didn’t share a typical GPU. Or, for that matter, widespread efficiency.
The RTX 4080 12GB, because it was, would have been primarily based on the smaller AD104 GPU, somewhat than the AD103 GPU used for the 16GB mannequin. In observe, this could have precipitated the 12GB mannequin to ship solely about 82% of the previous’s shader/tensor throughput, and simply 70% of the reminiscence bandwidth. A large efficiency hole that NVIDIA’s personal figures forward of the launch have all however confirmed.
NVIDIA, for its half, isn’t any stranger to overloading a product line on this vogue, with similarly-named elements delivering unequal efficiency and the distinction denoted solely by their VRAM capability. This was a observe that began with the GTX 1060 series, and continued with the RTX 3080 series. Nonetheless, the efficiency hole between the RTX 4080 elements was far bigger than something NVIDIA has beforehand executed, bringing a great deal extra consideration to the issues that come from having such disparate elements sharing a typical product title.
Of equal criticism has been NVIDIA’s resolution to promote an AD104 half as an RTX 4080 card to start with. Historically in NVIDIA’s product stack, the subsequent card beneath the xx80 card is a few type of xx70 card. And whereas video card names and GPU identifiers are basically arbitrary, NVIDIA’s early efficiency figures painted an image of a card that might have carried out quite a bit just like the form of card most individuals would count on from the RTX 4070 – delivering efficiency upwards of 20% (or extra) behind the higher RTX 4080, and on-par with the last-generation flagship, the RTX 3090 Ti. In different phrases, there was quite a lot of suspicion throughout the fanatic neighborhood that NVIDIA was making an attempt to promote what in any other case would have been the RTX 4070 as an RTX 4080, whereas carrying a better worth to match.
In any case, these plans are actually formally scuttled. No matter NVIDIA has deliberate for his or her AD104-based RTX 40 sequence card is one thing solely the corporate is aware of presently. In the meantime come November 16th when the RTX 4080 sequence launches, the 16GB AD103-based playing cards would be the solely choices accessible, with costs beginning at $1199.