McG
Datum registracije: Feb 2014
Lokacija: Varaždin
Postovi: 8,308
|
Flagship nVidia “Ada” SKU might feature 18176 cores, 48GB memory and 800W TDP
Citiraj:
Rumour has it that nVidia's working on a flagship "Ada Lovelace" graphics card that sits beyond nVidia's planned RTX 4090 graphics card. This GPU is reportedly using a new board design called PG137, and could be a next-generation RTX TITAN series graphics card. According to the leaker @kopite7kimi, nVidia's flagship next-generation CPU (called "The Beast") will feature 48GB of 24Gbps GDDR6X memory, 18,176 CUDA cores, and a TDP of 800 watts. This gives nVidia's so-called "The Beast" GPU two times as much VRAM capacity as nVidia's RTX 3090 Ti, memory modules that are 14.2% faster, and 69% more CUDA cores. With a TDP of 800 watts, nVidia's "The Beast" GPU will likely ship with two 12+4-pin PCIe 5.0 power connectors.
Today, the only graphics cards to do this are ultra-high-end overclocking-oriented graphics cards like EVGA's RTX 3090 Ti KINGPIN. At this time, it is unknown if this rumoured GPU model will be. It could be an RTX 4090 Ti, a new RTX TITAN, or something completely new from nVidia. Beyond that, it remains to be seen if nVidia is crazy enough to even release a graphics card with a TDP that's approaching a kW. As always, GPU rumours should always be taken with a grain of salt. While these rumoured GPU specifications are impressive, they sitting within crazy territory in terms of both power consumption and VRAM capacity. Cooling an 800W graphics card would be a very challenging task, and not many workloads can justify using having 48GB of VRAM. That said, every TITAN-class card from nVidia has always arguably had too much VRAM at launch.
|
Izvor: OC3D
EDIT:
Citiraj:
nVidia GeForce RTX 4090 rumored to launch in 450W & 600W variants, priced lower than RTX 3090 Ti
Citiraj:
The rumor states that the nVidia GeForce RTX 4090 might be the only "Ada Lovelace" powered graphics card launching and hitting retail shelves this year with the other variants possibly slipping into Q1 2023. One reason for this delay could be the fact that nVidia is very serious about its existing GPU inventory and AIBs could have asked the green team to wait a bit more before launching new cards & that is also why we are seeing the existing flagship lineup drop down the $1000 US price range in retail sites. nVidia's GeForce RTX 3080 Ti which had an MSRP of $1200 US is already being sold for lower than $800 US while the RTX 3090 which had an MSRP of $1500 US is currently on offer for $1000 US.
The more interesting part is the TGP figures. It is stated that the nVidia GeForce RTX 4090 graphics card will come in both 450W and 600W TGP flavors. The reference design is most likely going to be kept at 450W but AIBs will have the option to go up to 600W in their custom designs. Furthermore, it is clarified that the 450W and 600W variants will have the same frequency (boost clock) limit but the higher TGP variant will have a lower chance of hitting its TGP wall and may end up providing a slightly better performance uplift within games. nVidia's AIB partners will mention if the card uses the 450W or 600W TGP figure on their box packages which is something that we haven't seen in the last few generations.
|
Izvor: Interneti
|
Citiraj:
nVidia RTX 4070 Specs: 7,680 Cores, 12GB GDDR6X Memory, as Fast as the RTX 3090 Ti @ 300W
Citiraj:
The RTX 4070 will reportedly leverage 12GB of high-speed GDDR6X memory clocked at 21Gbps alongside the full-fat AD104 die. Featuring 7,680 FP32 cores across 60 SMs, 30 TPCs, and 5 GPCs, we’re possibly looking at at least 35-40 TFLOPs of single-precision performance. Since we’re talking about the fully enabled die, the L2 cache should also be untouched at 48MB. The RTX 4070 will be the first RTX x70 card to feature GDDR6X memory, a technology NVIDIA has limited to the $500+ segments. However, at the same time, the bus width has also been reduced to 160-bit keeping the bandwidth largely the same. The TGP is also a very respectable 300W, 80W more than its predecessor, the RTX 3070. Keep in mind that we’re looking at RTX 3090 to 3090 Ti levels of performance in a $400-500 SKU here, so this is nothing to scoff at.
|
Izvor: Hardware Times
|
Citiraj:
nVidia GeForce RTX 4080 Graphics Card Now Reportedly Features 9728 Cores, 16 GB GDDR6X Memory, 420W TBP, 35% Faster Than RTX 3090 Ti
Citiraj:
The nVidia GeForce RTX 4080 is expected to utilize a cut-down AD103-300 GPU configuration with 9,728 cores or 76 SMs enabled of the total 84 units whereas the previous configuration offered 80 SMs or 10,240 cores. While the full GPU comes packed with 64 MB of L2 cache and up to 224 ROPs, the RTX 4080 might end up with 48 MB of L2 cache and lower ROPs too due to its cut-down design. As for memory specs, the GeForce RTX 4080 is expected to rock 16 GB GDDR6X capacities that are said to be adjusted at 21 Gbps speeds across a 256-bit bus interface. This will provide up to 672 GB/s of bandwidth. For power, the TBP is said to be rated at around 420W which is close to the RTX 4090 and since the memory clocks aren't that aggressive, it is likely that the core clocks could be tuned for maximum graphics performance.
|
Izvor: Interneti
|
__________________
AMD Ryzen 9 9950X | Noctua NH-U12A chromax.black | MSI MAG B650 Tomahawk Wi-Fi | 128GB Kingston FURY Beast DDR5-5200 | 256GB AData SX8200 Pro NVMe | 2x4TB WD Red Plus | Fractal Define 7 Compact | Seasonic GX-750
AMD Ryzen 5 7600 | Noctua NH-U12A chromax.black | MSI MAG B650 Tomahawk Wi-Fi | 128GB Kingston FURY Beast DDR5-5200 | 256GB AData SX8200 Pro NVMe | 2x12TB WD Red Plus | Fractal Define 7 Compact | eVGA 650 B5
Zadnje izmijenjeno od: The Exiled. 05.08.2022. u 13:08.
|