
The NVIDIA Quadro RTX (Ampere) will feature a GA102 chip with 48 GB of GDDR6. Due to GDDR6X shortages, its bandwidth will be lower than the RTX 3090.
Less than a week ago, NVIDIA unveiled its new line of consumer GeForce RTX 30 graphics cards. With their specs already thoroughly dissected, insiders and tech media have shifted their focus to the yet-unannounced Quadro RTX series refresh. Especially since a new leak reveals intriguing details about the top-tier Quadro product.
⤢ ВІДКРИТИAccording to Moore's Law Is Dead, the most powerful NVIDIA Quadro RTX graphics card of the Ampere generation will feature a full-fledged GA102 graphics chip with 10752 CUDA cores and a 48 GB GDDR6 video buffer.
Apparently, GDDR6X memory manufacturers currently can't provide the required 48 GB capacity, so NVIDIA simply had to equip its flagship product with an older memory type.
By loading, you allow content from X (Twitter) (cookies may be set).
Consequently, the bandwidth of the NVIDIA Quadro RTX variant that surfaced is 768 GB/s, which, as you might expect, is significantly lower than the NVIDIA GeForce RTX 3090's impressive 936 GB/s.
Most likely, next year, when the production of suitable high-capacity GDDR6X dies ramps up, the green giant will release an updated Ampere-generation NVIDIA Quadro RTX with advanced video memory onboard.
Source: WccfTech