Rumor Mill: Nvidia Prepping Flagship RTX 3090 Ti GPU
If there’s one thing the gaming world needs right now it’s another outrageously expensive and unobtainable graphics card, and Nvidia is heeding the call with reports of an alleged top-tier GPU waiting in the wings. Dubbed the RTX 3090 Ti, this full-blown Ampere card will offer the entirety of the GA102 die’s performance envelope, along with higher clocked memory from Micron to help distance the card from its lowly RTX 3090 predecessor.
The rumor springs forth from Twitter user Uniko’s hardware, who tweets the new GPU will feature new 21Gb/s memory from Micron, which is a bit of a boost from the memory used in the current RTX 3090, which runs at approximately 19.5Gb/s. The memory chips also feature twice the capacity of the previous modules, so only half the number will be required to reach the allotted 24GB of GDDR6X in the new GPU. This reduction in memory chips should allow the card to run a bit cooler, despite those chips requiring more power than the previous model due to the higher clock speeds. The GPU will also keep the same 384-bit bus as the current card, allowing it to theoretically offer up to 1TB/s of memory bandwidth.
Other salient specs include the full allotment of GA102’s horsepower, including 10,752 CUDA cores (up from 10,496), the aforementioned 24GB of a super-fast memory, and a total board power of around 400w or so, which is 50w more than the current RTX 3090. That’s an increase in core count of 2.5 percent, and when you throw in the faster memory, it seems wise to assume that overall the new GPU will be about five percent faster than the previous GPU. There is no information at this time about reported clock speeds, however. WCCFtech is reporting that despite being more powerful than the card it replaces, MSRP should remain the same at $1,499.
If this card does indeed exist, its arrival follows a pattern established by Nvidia in the past, where it released a cut-down version of its biggest die first, then follows up with an unfettered “big” version of the chip at the end of the product’s lifecycle, though previously it was branded Titan. Nvidia seems to have abandoned the Titan naming altogether for some reason, and is now swapping back-and-forth between Super and Ti for upgraded versions of its current GPUs. While it favored Super in the previous Turing era, it has switched to Ti for its Ampere upgrades.
Still, the news of this card’s imminent release begs the question, “Why?” We already have an RTX 3080 Ti, which is very close in specs to the RTX 3090, aside from having half the memory. Not to mention the fact that neither of these GPUs can’t be purchased for anywhere near their MSRP, leaving them costing over $2,000 on third-party sites, assuming you can even find one for sale.
Regardless, this rumored GPU is supposed to break cover in January 2022, perhaps as the company’s big announcement for CES. Go talk to your loan officer now, and watch this space.
Continue reading
Nvidia Reportedly Prepping Two New RTX 3050s, One with Just 4GB of VRAM
The company seems intent on giving gamers, and miners, some options in the sub-$300 space.
Nvidia May Be Prepping a New GTX 1050 Ti Max-Q to Counter Intel, AMD
Nvidia is reportedly readying a new version of the GTX 1050 Ti with a Max-Q spin on the silicon. Is it a move against AMD's just-announced Vega Mobile or a shot across the bow of Intel's Vega-equipped CPUs?
Apple Prepping New Coprocessors, Macs, iPad for Later This Year
Apple is already hard at work on a new set of chips and hardware for this year, but Meltdown and Spectre could have ramifications well beyond Intel.
Sega Is Prepping a Mega Drive Mini, but Who’s Building It?
Sega is bringing out a new Mega Drive Mini to celebrate its 30th anniversary, but we don't know who is building the hardware yet — and given ATGames' mixed track record, that matters.