Nvidia’s RTX 4080 graphics card could be a tiny significantly less beefy than has been previously rumored, which follows quickly on the heels of news that the RTX 4070 could be a peppier GPU than predicted.
This fresh new Lovelace speculation once more arrives from Kopite7kimi, a normal hardware leaker on Twitter who has been supplying rumors at really a pace of late (while that’s to be expected as the RTX 4000 variety nears its launch timeframe, with an envisioned debut in or all around Oct, assuming no delays).
I’m not a chatterbox, but I have to make some updates. I hope you do not intellect. a feasible RTX 4080, PG136/139-SKU360Ad103-300-A19728FP32256bit 16G 21Gbps GDDR6Xtotal electrical power ~420WTSE ~15000Now I have done the latest update for 4090, 4080 and 4070.August 5, 2022
The idea is that the ‘possible’ RTX 4080 – a reminder that practically nothing is necessarily established in stone, even if this is Nvidia’s existing wondering with the spec of the GPU – could operate with 9,728 CUDA Cores, which is less than the beforehand rumored 10,240.
The leaker thinks that the RTX 4080 will appear outfitted with 16GB of GDDR6X (at 21Gbps), and a ability use of all around 420W, all of which stays in line with what was floated beforehand – the only change is that trimming down of the main count at this point.
Evaluation: Nvidia’s positioning of RTX 4000 designs
The to start with thing to take note below is that this is not a big modify. Dropping down 512 CUDA Cores – which usually means reducing the streaming multiprocessors on-board by 4, from 80 to 76 – is not a massive shift, but it does however signify a slight reduction in uncooked grunt for the GPU.
Let us bear in mind that this is just a rumor, and the leak landscape close to Lovelace is seemingly shifting really consistently these days – which might be a reflection of Nvidia tinkering with and changing relative specs as the company moves toward a finalized bunch of products.
As talked about at the outset, Kopite7kimi also just lately brought us information on the RTX 4070, with the apparent notion for that GPU currently being to up its functionality amount at this stage – to the spec that was formerly rumored for the RTX 4070 Ti, in truth. A quite massive stage up, for positive. So, assuming this is all correct, what may well Nvidia’s reasoning be in this article?
Bringing down the CUDA Core rely of the RTX 4080 a little bit, while severely beefing up the RTX 4070 – not just for CUDA Cores, but the loadout of VRAM as effectively – implies, clearly more than enough, that the relative general performance of the RTX 4070 and RTX 4080 is theoretically remaining introduced additional closely in line.
With much less differentiation in between the two, and Nvidia’s xx80 model commonly being a excellent chunk much more costly than the far more mainstream xx70 card – the comparatively reasonably priced higher-conclude GPU, if you will – with Lovelace, are we going to be seeking at an RTX 4070 which is even far more in the sweet spot for most consumers?
Or to put it a different way, if the RTX 4080 isn’t all that considerably faster, who will fork out for just one somewhat than a 4070? Except inventory is an difficulty for the latter, of study course that could occur into perform, notably if it is genuinely well known.
Also, we require to remember that CUDA Cores aren’t the whole story for functionality, with quite a few other variables coming into play, like clock speeds. It is right here that there could be a clue in the RTX 4070 supposedly sticking with around 300W for electricity use, while the RTX 4080 will purportedly hit 420W or thereabouts – this could point to Nvidia severely beefing up clock speeds (and hence efficiency) with the larger-finish GPU when compared to the 4070.
Additionally, Kopite7kimi has taken care of the past guestimate for a 3DMark TimeSpy Extraordinary rating for the RTX 4080 (all-around 15,000), in spite of those CUDA Cores currently being shaved off, so apparently plenty of, that could also advise Nvidia could possibly be pushing tougher with clocks here for the 4080. Despite the fact that the leaker does not mention clock speeds precisely.
Absolutely it doesn’t make perception to have the RTX 4080 as a somewhat unattractive proposition compared to the 4070 – a a great deal bigger electrical power usage won’t enable either, in conditions of forcing PSU updates, and electricity prices in standard – so for now, we’ll assume that if this rumor is true, we just can’t yet see the full picture of what Nvidia is striving to do with the positioning of Lovelace GPUs. Or that there’ll be a different rejigging of specs however to come…
Relative pricing could be a factor listed here, much too – we haven’t read a lot on that front from the rumor mill nonetheless, but we cannot visualize the RTX 4000 series will be pitched with affordability in thoughts. Specifically not if Nvidia has a large amount of surplus RTX 3000 inventory to deal with, as the grapevine contends, and shifting all that overlaps with the Lovelace start (which is very likely only a few of months away now).
Whatever the case, you’d envision Team Inexperienced would’ve acquired from its mistakes in conditions of the RTX 3080 getting off on the improper foot, coming with only 10GB of VRAM in its original incarnation – an unpopular go compounded by the later on launch of a 12GB variant. The latter was beefier not just in phrases of the added movie memory, leaving some customers of the original RTX 3080 with a severe situation of remorse.