Thursday, December 26th 2024
NVIDIA GeForce RTX 5090 Powered by "GB202" Silicon, 512-bit GDDR7, ASIC Pictured
Here is the first picture of what is very likely the GeForce RTX 5090 "Blackwell," the successor to the RTX 4090 "Ada." The picture, its massive GPU, and layout appear to confirm the weekend's bare PCB leak. The RTX 5090 is powered by the "GB202" silicon, the largest gaming GPU based on the "Blackwell" graphics architecture. The silicon in the picture has the ASIC code "GB202-300-A1." From this ASIC code, we can deduce that the RTX 5090 may not max out the silicon (i.e. enable all SM present on it), as maxed-out NVIDIA ASICs tend to have the variant designation "450."
The "GB202" ASIC is surrounded by sixteen GDDR7 memory chips, which reportedly make the 32 GB memory size of the RTX 5090. The chip count, coupled with the large GPU package size (high pin-count), confirm that the "GB202" features a 512-bit wide memory bus. Assuming a memory speed of 28 Gbps, this memory bus should yield a stellar memory bandwidth of 1,792 GB/s. The GPU and memory are surrounded by the card's 24-phase VRM solution. This draws power from a single 16-pin 12V-2x6 power connector. NVIDIA will likely max out the 600 W continuous power-delivery capability of the connector, and give the card a TGP of around 500-550 W, if not more.
Source:
harukaze5719 (Twitter)
The "GB202" ASIC is surrounded by sixteen GDDR7 memory chips, which reportedly make the 32 GB memory size of the RTX 5090. The chip count, coupled with the large GPU package size (high pin-count), confirm that the "GB202" features a 512-bit wide memory bus. Assuming a memory speed of 28 Gbps, this memory bus should yield a stellar memory bandwidth of 1,792 GB/s. The GPU and memory are surrounded by the card's 24-phase VRM solution. This draws power from a single 16-pin 12V-2x6 power connector. NVIDIA will likely max out the 600 W continuous power-delivery capability of the connector, and give the card a TGP of around 500-550 W, if not more.
36 Comments on NVIDIA GeForce RTX 5090 Powered by "GB202" Silicon, 512-bit GDDR7, ASIC Pictured
It basically needs the same PSU as this 350w card.
Just enlight me about the cannot exceed its power rating statement also. Please explain.
Anyway, this GPU hyper performance will be only matched by its unattainable cost.
All the way back to the 8800gtx nvidia was stating CUDA was their future. When the first Titan hit the selling point was it's use in professional instances it wasn't until companies like Falcon NW decided it was SLI in a card slammed into an ITX case and people like Linus started buying them up that it being a "gaming" card took hold.
GeForce doesn't even really mean gaming. Tons of companies deploy GeForce based laptops that will never game but use the GPUs for other professional things that do not need the Quadro drivers or the price associated with those.
Reintroduce that tier of card as a XX90 and start the price lower, then gradually ratchet it up.
Nvidia already hard locked the power draw, spikes wise, on the 4090 when you bring the PL to 600w.
Nvidia barely took advantage of the pci-express slot power delivery on the RTX 4000 serie, they could bring it back with an additional ~50w contribute from it. But it isn't a game changer and they probably stopped using it as a safety measure.