We’ve got a date for Nvidia’s subsequent main GPU dump. It is third of January subsequent 12 months and a GeForce Particular Occasion keynote for the CES tech jamboree in Las Vegas (opens in new tab), albeit Nvidia shall be conducting the occasion remotely.
Past describing the occasion as detailing “the newest applied sciences, merchandise and software program from GeForce”, Nvidia is not getting extra particular. Nonetheless, we’re anticipating to see two principal bulletins on the occasion.
On the desktop aspect, very possible we’ll lastly see what Nvidia has accomplished with the cancelled RTX 4080 12GB mannequin (opens in new tab). Odds are it will likely be rebranded RTX 4070 Ti and be introduced on the third at that CES occasion.
Given the RTX 4080 12GB boards had been already in mass manufacturing when Nvidia had a change of coronary heart, we’re not anticipating a lot by means of a change in specs within the transition to RTX 4070 Ti branding.
So, that is 7,680 shaders, aka CUDA cores in Nvidia parlance, 240 texture models, 80 ROPs and 12GB of reminiscence operating over a 196-bit bus. Roughly talking, these numbers are round 40 to 45 p.c of the flagship RTX 4090 GPU (opens in new tab). Whereas it is attainable Nvidia could tweak a few of these specs for the purported RTX 4070 Ti by way of a video BIOS replace, we’re not anticipating a dramatic deviation.
As a substitute, the large unknown is pricing. The RTX 4080 12GB was priced at $899 and the belief is that the RTX 4070 Ti shall be cheaper. However how less expensive? Odds are, we’ll know on January 3.
The opposite main announcement on the CES occasion is anticipated to be cell variants of Nvidia’s RTX 40-series GPUs, in any other case often known as Ada Lovelace. We’re anticipating to see RTX 4080 and 4070 branded cell GPUs introduced at minimal, with the attainable addition of a cell RTX 4090 and 4060 cell chips to prime and tail the vary.
Primarily based on the perfect at present out there info, the cell lineup would subsequently look one thing like this. On the very prime, the RTX 4090 will use the identical AD103 GPU because the desktop 4080 card and pack 9,728 shaders, 304 texture models and 112 ROPs, all operating over a 256-bit reminiscence bus. Clock speeds can be considerably decrease than the desktop 4080 at roughly 1.4GHz.
The RTX 4080 and 4080 Ti cell will step right down to the AD104 GPU briefly seen in that cancelled 4080 12GB desktop board, with the Ti mannequin providing a lot the identical {hardware} options because the 4080 12GB desktop however operating at decrease clock speeds. Subtract roughly 15 p.c from the shader, texture and ROP counts to reach on the non-Ti mannequin. The smaller AD104 chip ought to clock somewhat larger in a cell implementation and would possibly flirt with as a lot as 2GHz in larger performing installations. Nonetheless, the same old lower-power Max-Q variants for thin-and-light will inevitably be a lot slower.
As for the RTX 4070 cell, that is more likely to be primarily based on the as-yet unannounced AD106 chip. Specs will possible fall round 4,608 shaders, 144 textures and 48 ROPs, plus a 128-bit reminiscence bus. Lastly, the RTX 4060 might use one other unannounced chip, AD107, with additional cut-down specs, more likely to be round 2,560 to three,072 shaders, 80 to 96 texture models and round 32 ROPs. Mercifully, AD107 is at present anticipated to stay with a 128-bit reminiscence bus reasonably than stepping down even additional to a 96-bit bus.
So, there you’ve it: what to anticipate on January third. Tell us beneath whether or not that sounds thrilling or if it simply provides to the frustration on the newest GPU launches.
